Skip to main content

November 2026

November 16, 2026

API
  • ⚠️ Logistics: The v2 leg record now follows the v2 contract. Review any client that reads or writes v2 legs.
  • ⚠️ Logistics: The GLEC distance adjustment now applies in every channel and is derived from the client’s distance (client_distance_km). Review any client that sends or reads the adjustment.
  • DMA:
    • Survey results are now available per response, with reasons.
    • New IROs can only be created under shortlisted topics.
    • You can now reopen a submitted survey and read a stakeholder’s history.
  • Logistics:
    • Subcontracted legs are now computed at read time, rolled out per root organization.
    • GLEC shipments now ask the same questions in the sheet, the template and the PDF review.
    • The GLEC recharges template now asks for the unit of the quantity, which must be the fuel’s unit.
    • There is now one hub per normalized name, matched the same way everywhere.
  • ISO 14001: You can now bulk upload the legal register’s requirements that were written by hand.
  • Organizations:
    • An organization can now have several internal codes, unique within the group.
    • Internal code and notes are now matched by every import Organization column.
  • Imports:
    • Datasets and project tasks now cite the spreadsheet they were imported from.
    • ISO risks and opportunities with no project are now imported as the organization’s.
    • Remembered organization, facility, country and currency values are now shared across templates.
    • Table recipes can now be built from the fixed values people typed on earlier deliveries.
    • The wastes pair, future date, disabled facility and corrected rows rules now decide alike in the wizard and the rules catalogue.
  • SRA: You can now accept Leia’s proposals for many suppliers at once, and the answer picked for a mention.
  • Supplier risk: The suppliers list now returns each answered question’s state and options.
  • Risk matrix: The organization risk matrix list can now be filtered by a row’s source.
  • Management systems:
    • Market countries are now validated.
    • Objective status is now an enum type.
    • The OCR category is now nullable.
  • Invoices: Saving a PDF review now assigns the supplier per facility country.
  • Reviews: The custom columns typed in any document review are now saved with its records.
  • Auth: Sign-ups from the larger disposable email list are now refused, unless the mail is on Google or Microsoft.
  • Fix: Facilities no longer blocks a facility move when a group upload is owned by an ancestor of the target.
  • Fix: Auto-confirm now keeps a document for review when a required custom column is saved by the column editor.
  • Fix: Ingestions processing feed now lists files only, never a finished import’s job.
  • Fix: Energy now rebuilds the default energy configs when you recalculate energy.
  • Fix: Imports can now delete a row from any API process, even if it was added on another.
  • Fix: Imports wastes pair rule now only checks years an activity can have.

November 15, 2026

API
  • Upload cycle:
    • Heat bills and fuel tickets are now auto-confirmed.
    • A vehicle’s page now fixes its vehicle on the upload.
    • A zero quantity is now accepted.
    • A packaging sheet can now be re-read.
  • Review: A reviewed document now saves its custom columns and its attachment. Required columns now wait for a person to fill them in.

November 14, 2026

API
  • Upload cycle: PPWR suppliers and SKUs, and a crop’s inputs and farms, now go through the shared upload flow.
  • Use of product: A change of emission factor now shows in the history of the use.

November 13, 2026

API · MCP
  • MCP: You can now create, edit and delete sub-sections of the SRA risk model from the MCP tools.
  • Supplier risk:
    • You can now run a custom import from Purchases, filtered by period, amounts, country, name, expense type and code.
    • A supplier imported from Purchases now keeps a link to the Purchases supplier it comes from.
  • Imports:
    • Tab names now fill the field they fit, such as a month, a site or a company.
    • A submitted session’s rows can now be read while its parquet file is being stored.
  • Ingestions: The detail of a tabular ingestion now names the processing job that saves its rows.
  • Upload cycle:
    • Page contexts, tab templates and the form’s rules now apply in every upload entry point.
    • Each record now shows where it comes from. A single split rule is now used. The review uses its own emission factor.
  • Supplier pre-evaluation: Failed Leia analyses now give a clear failure reason. You can now retry a single supplier. Engine fixes are included.
  • Fix: Imports now import a fleet larger than the value step reads, instead of holding it.
  • Fix: SRA can now resend a questionnaire when none exists. Leia is now confirmed on the server before submit.
  • Fix: Custom KPI evidence files no longer show as failed processing in My files.

November 12, 2026

API
  • Imports:
    • Import sessions now say which columns the upload page fixes or defaults.
    • A workbook filled in ahead of time now imports only what has been measured. A file with rows left out now says so.
    • Waste imports now give each line its own confidence. NACE and emission-factor column headers are now recognised in 7 languages.
  • Supplier risk: You can now import the Purchases suppliers that make up 80% of the spend or of the CO2e.
  • SRA: The risk questionnaire now has sub-sections.
  • LCA: Ecoinvent matching for materials now uses a material grade table.
  • Fix: Invoices no longer list a facility’s supply point twice when invoices add their CUPS.
  • Fix: Employees survey-rows list now supports the name is empty and not empty filters.
  • Fix: Reports created by a member now cover the member who authored them, not the whole group.
  • Fix: Bulk edits in the upload cycle are now typed and versioned. Unlinking a record from the trash now works when it has no author.

November 11, 2026

API · MCP
  • MCP:
    • get_water_consumption now reads water in m³ the same way the water charts do. It works for a business group or for a list of its organizations.
    • You can now create and edit the supplier risk model and its Leia rules from the MCP tools.
    • get_file now returns import_state, and its guidance follows that state.
    • Pre-registered OAuth clients are now available for Microsoft 365 Copilot connectors.
  • Imports:
    • Several tabs of a workbook can now be read as one table. The sheet step suggests which tabs to combine and answers for them as a single table.
    • The auto-import job reads several tabs as one table when the planner or a recipe is sure they belong together. The planner decides on each tab’s part or not-part call.
    • The auto-import job now signs its session at the start and records every stop on it.
  • Social data: Social and governance answers can now be uploaded in bulk, together with their source.
  • Datadis: Supply contracts can now be bulk-updated, either by a list of ids or by the list’s filters.
  • Supplier pre-evaluation: Bulk runs with many items are more robust. Evaluations now take the supplier’s country into account and record the provenance of their evidence.
  • Invitations: An invitation to an address that could never sign up is now refused.
  • Confidence: A field typed or changed by hand now scores 1.0 in the record’s reading confidence.
  • Fix: Ingestions now give every record of a document the confidence of its own item.
  • Fix: Imports now derive a spreadsheet’s import_state consistently:
    • import_state reads the auto-import job’s transport and bounds the submitting state.
    • It reports that Leia works on a spreadsheet from the moment its session is created.
    • The spreadsheet’s current_step, tab and auto_import now follow import_state.
    • The detail view is built from one session.
  • Fix: Imports now give a reused ecoinvent link the region of the new country in the LCA materials import.
  • Fix: Purchases v2 supplier create and update now accept every SIC code.
  • Fix: Vehicles emission-factor matching is cleaned up, and the FINAL fallback now takes the year into account.
  • Fix: Pre-audit now flags trucks in the transport distance check only beyond 5,000 km.
  • Fix: Pre-audit no longer raises false CRITICAL findings for items that the validated views already count.

November 10, 2026

API
  • ⚠️ Invoices: Heat invoices now require stationary_fuel_id, so that their kWh can be calculated.
  • Supplier pre-evaluation: Leia can now analyse suppliers in bulk. Runs from the same bulk analysis are grouped by batch_id.
  • Imports:
    • Import sessions now expose blocked_reason and auto_started_at. The auto-import job writes both on its session.
    • An auto-import session is now created with its schedule and unread. For an xlsx file, it is created at classification, before the sheet is read.
    • A file:import_state event is now emitted after every committed change to a file’s import session.
    • Import sessions no longer expire. Their artifacts are removed together with the session, file, project or organization.
    • Many more import templates now run on the snapshot session protocol. They cover ISO registers, company strategy, processes, crops, SFDR, waste water, events, project tasks, PPWR, SRA, LCA scenario materials, use of product, elastic datasets, supplier codes and survey emails. A schema snapshot is also backfilled onto open legacy import sessions.
    • The value model now reviews category picks that the text check refused. A row of spaces is no longer treated as data.
  • Ingestions: Spreadsheets now report an import_state, derived from their three layers.
  • Record sheet:
    • Product periods can now be updated with PATCH.
    • PPWR now carries a confidence value.
    • The v2 workforce record now supports attachments.
  • Land management: Farms and crop inputs now support attachments.
  • Invoices: Dismissals of consumption issues are now stored. Ignore and a confirmed 0 now persist.
  • Datadis: You can now list the linked NIFs across the whole business group.
  • Fix: Datadis now handles a moved contract end date by deleting the invoices’ impacts and stopping the sync.
  • Fix: Ingestions now read a waste document with a single date as a one-day period on that date.
  • Fix: Ingestions now read the received net kg on a waste DI as the waste quantity.
  • Fix: Pre-audit no longer flags GdO redemption offsets as orphan electricity transactions.
  • Docs: The v2 attachment docs now state that attachments are create-only.

November 9, 2026

API · MCP
  • MCP: The MCP tools can now reach the MultiDB emission-factor model for wastes and other categories.
  • Fix: Imports now offer Waste and Processes in the project-tasks category list.
  • Fix: Vehicles can now be edited when the stored plate has stray whitespace. The edit no longer fails as a duplicate plate.
  • Fix: Ingestions can read pre-cutover vehicle fuel PDFs natively again.
  • Fix: Purchases now report a declined description match, so the detail can say the supplier factor was used.

November 8, 2026

API
  • Invitations: The invitation link now carries a token and an expiry.
  • Supplier pre-evaluation: Engine quality, cost and time have improved. A limit of 70 pre-evaluations per day per payer now applies.
  • Docs: The v2 Logistics hubs and delete pages now match the code.

November 7, 2026

API · MCP
  • ⚠️ Legacy API: The legacy GET /api/v1/custom_emission_groups endpoint was broken and has been removed. The OpenAPI dump script is also fixed.
  • MCP: The new upload_file_content tool uploads a file’s content for clients that cannot send the presigned PUT.
  • Supplier pre-evaluation: Leia now applies a rule per risk-model question. The engine is optimised, and items that can’t be verified are now reported as unverifiable.
  • Supplier answers: Leia now proposes an answer for each question. You can accept or reject each proposal, and the score is computed from the accepted answers.
  • Imports:
    • PATCH /rows can now set values on every row of a match.
    • /status now returns what a resumed wizard needs.
    • /validate now accepts the wizard’s context.
    • /validate without value_mappings keeps the saved mappings. Sending {} clears them.
  • Auth: Sign-ups from domains on the public disposable-domain list are now refused.
  • Users: Deleting a user no longer deletes business data.
  • Fix: Imports PATCH /rows now keeps the errors of the cells it doesn’t correct.
  • Fix: Imports PATCH /rows now corrects only a validated session and returns a clear error otherwise.
  • Fix: Imports keep a cell’s error when a number or category correction doesn’t coerce.
  • Fix: Imports now report a letter in a number cell as a format error instead of dropping it.
  • Fix: Imports now import a file’s sheet once, however many of its sessions are submitted.
  • Fix: Imports now hand a session back as validated when a submit is refused after its claim.
  • Docs: The v2 Logistics docs now describe the behavior the code implements.

November 6, 2026

API · CLI
  • ⚠️ Logistics: POST /v2/logistics/requests now requires subcontractor. Requests without it are rejected.
  • ⚠️ Imports: ensure-from-raw is deprecated. Use POST /v2/files/{file_id}/import-sessions instead.
  • Webhooks: file.finished is now sent for spreadsheets uploaded with an API key.
  • Employees: POST /v2/employees accepts on_existing=reuse.
  • LCA:
    • A new endpoint returns ecoinvent activity detail.
    • Lookups now prefer the country’s “market for” activity over “market group for”.
  • Imports:
    • Category-page uploads now go through the auto-import job, and a file is marked imported only after its rows are saved.
    • A project’s tasks spreadsheet can now be uploaded through the shared upload flow.
  • Files: A document that Leia couldn’t process is now kept as evidence only.
  • Record sheets: The Attachment and Confianza columns are now available across the remaining record sheets.
  • Source of truth (SoT):
    • The list endpoints now return answered.
    • A widget can now answer a calculated datapoint.
    • Sending an empty figure now clears the datapoint.
  • Purchases: An ERP code is now read again together with its supplier and buyer, before falling back to the supplier’s sector.
  • Logistics: Own-fleet intensity values are now recomputed every night.
  • Fix: CLI uploads now send the signed headers, and auto-imported files reach the files worker.
  • Fix: Purchases now reject a supplier country that isn’t an ISO-2 code before creating the supplier. Supplier countries are stored as ISO-2 codes, and legacy values are read without failing.
  • Fix: Imports no longer fail an upload because an .xlsx file has formula error cells such as #DIV/0!.
  • Fix: Imports now explain why a spreadsheet is unreadable, instead of reporting “unsupported format or encoding”.
  • Fix: Imports now hold the auto-import job with owner_cannot_write when the file’s owner has no write access.
  • Fix: Files can no longer be moved to another importer once a spreadsheet has been imported.
  • Fix: Files deleted after import now remove the absences, trainings, commutes and fuel recharges they created. The import is also undone when no row was saved correctly.
  • Fix: Files batch delete now returns 409 FILE_USED_BY_ANALYSIS for the source file of a reporting suggestions run.
  • Fix: Reporting suggestions runs whose organization leaves the project now end as project_unavailable.
  • Fix: Reports widgets now use fiscal years when the report period is a fiscal year.
  • Fix: Reports for Scope 1 and 2 now show the record’s own fuel instead of the deprecated facility_fuel.
  • Fix: Scope report status checks now apportion invoices that straddle the report window.
  • Fix: Scope report v2 now fails on a database error instead of returning an incomplete report.
  • Fix: Logistics now seeds a new TOC’s intensity value for the leg’s year.
  • Docs: The OpenAPI spec now describes naive datetimes and problem-format 422 responses as the API sends them.

November 5, 2026

API · CLI · MCP
  • ⚠️ Request size limit: The API now refuses request bodies over 100 MB. The spreadsheet readers used by the document classifier are also guarded against malformed workbooks.
  • ⚠️ Invalid periods: A record whose period ends before it starts now fails instead of passing with no impacts.
  • Logistics: GET /v2/logistics/packages reads up to 100 packages by id, each with its current totals.
  • CLI and MCP: An agent can now upload a file and follow its processing, using routes that are already live.
  • Import API: The Import API’s routes now work on a file.
  • Purchases: Every channel now returns the same data and uses the same method. Purchases expose their full history and can be filtered by those that need review.
  • Vehicle consumptions:
    • You can recalculate selected vehicle consumptions by activity category.
    • A consumption read from a ticket by Leia now keeps the confidence of that reading.
  • Management systems: The organization’s objectives now have an Activity and an Attachment.
  • Double materiality assessment (DMA): The stakeholder engagement now has a history covering stakeholders, surveys, questions, links and answers.
  • SASB: Answers can now be tables. A question counts as answered when it has a narrative, a figure or a table. The catalogue is now v2, and the wizard has lookups.
  • Imports:
    • ISO objectives imported with no project now belong to the organization.
    • A new upload no longer deletes the template’s other open sessions.
  • Files: The shared upload flow inside a project now links what it creates to that project.
  • Document processing:
    • Procesos PDFs are read as process-emission invoices, each line with its own factor.
    • Plant reports (EDAR) uploaded from Tratamiento de aguas are read as one measurement per plant and period.
  • Fix: Vehicle consumptions bulk uploads now say why a row failed, and the preview catches undeclared fuels.
  • Fix: Vehicle consumptions no longer put the review’s line key in custom_id.
  • Fix: The v2 vehicle class-fuel lookup is now filtered by fuel.
  • Fix: Imports no longer suggest an empty facility column as mapped.
  • Fix: EcoVadis now treats a document that is valid today but lapses before the due date as a warning, not as expired.
  • Fix: Logistics imports no longer time out on tabular detail.
  • Fix: Geographic maritime distances cached at 0 km are now recomputed.
  • Fix: DMA survey generation now reads the model’s final answer and lists short topic texts.

November 4, 2026

API
  • Record comments: Comments are now available on suppliers, vehicles and vehicle consumptions. Suppliers also expose their change history.

November 3, 2026

API
  • ⚠️ Activity dates: Every v2 activity date is tomorrow in UTC at the latest. Wastes, invoices, vehicle consumptions and hotel check-outs no longer take a later date; purchases, waste water treatments, accidents and birth dates now take tomorrow.
  • Deprecated: The consumption bulk deletes, the transport file uploads and POST /v2/employees/bulk-delete now point to their v2 routes and are retired on April 30, 2027. The deprecated pages say which cases still need them until then.
  • Fix: The deprecated single deletes now answer their Deprecation, Link and Sunset headers.
  • Free plan: Organizations on the free plan can now open and delete their webhook endpoints, as they list them.
  • Purchases: A purchase the free plan’s quota of description lines left on its supplier’s factor now says so with PURCHASE_DESCRIPTION_QUOTA_EXHAUSTED, or next to PURCHASE_CALCULATION_NO_TOTAL_IMPACTS_PRODUCED when a purchase with only non_currency_quantity fails.

November 2, 2026

API
  • Double materiality assessment (DMA): IRO identification now has a legacy flow. IRO assessment now builds on topic identification. Suggestions are now generated only in the language of the assessment, and the organization name is now used when generating them.

October 2026

October 30, 2026

API · MCP
  • Purchases: You can now attach a document to a purchase.
  • Fix: MCP SSO sessions no longer expire after 60 seconds. Clerk tokens minted server-side now authenticate without the azp claim.
  • Fix: DMA generations created by Dcycle are now restricted to organization admins in the API as well.
  • Fix: Remuneration totals are now the simple mean of their employees, matching the per-row values.
  • Fix: CDP now confirms a legacy prose proposal on a matrix question as text.

October 29, 2026

API
  • Business travel: MITECO can now be chosen as the emission factor database for cars and motorbikes.
  • Fix: Waste water treatments now treat a null recovered methane value as 0 in v2 and after a v1 edit.
  • Docs: The waste water treatments documentation now asks for chemical oxygen demand (COD) on industrial lines and for dry sludge.
  • ⚠️ Deprecated: Every route with a v2 successor is retired on April 30, 2027; its responses carry Sunset. The changelog lists each route and its successor.
  • Elastic data: Deleting a dataset shared with the subsidiaries deletes it in your organization only; each subsidiary keeps its copy with its records, and the delete answers which copies stayed.

October 28, 2026

API
  • Single project rule: Projects now resolve access through common resolvers, so every project-scoped route follows one access rule. The project-access route inventory is closed and the residual access code has been retired. Review any client that relies on project-scoped access behavior.
  • Projects: The owner pays and members propose. Project perimeter and consolidated output are now computed by the new project engine. Project dashboards, report answers, and taxonomy results use the same consolidated output.
  • Files: Project file links now have a contribute level, and common documents are supported.
  • Invoices: PUT /v1/invoices/{invoice_id} that swaps one facility for another, of your organization or of an organization below it, keeps the invoice’s id, comments, project assignments and custom-column values; a move to another organization answers 409 when it would lose data.
  • Suppliers: Suppliers are now owned per organization and use a common risk model. SRA and PPWR results are consolidated across the project.
  • Supplier risk assessment (SRA): Supplier lookups are now scoped by the contributing organization.
  • Decarbonization: Projects have a common plan, and each organization keeps its own initiatives. Leia now answers over the caller’s own data.
  • Reports: Reports now record their authorship and can link to LCA projects.
  • Mobility: Transport plans and logistics rows are owned by the organization that created them. Project figures are consolidated.
  • Management systems: ISO management systems now follow the common project rule. The configuring organization is now distinct from the contributing organization.
  • Auth: Signing out now ends every session immediately and is recorded.
  • Fix: An unlink can no longer revoke access while an invite accept is still being committed.
  • Fix: An opaque media key is now signed only for a reader of its owner.

October 27, 2026

API
  • ⚠️ Free trial plans: Organizations on a trial can no longer use modules that are clearly paid: EU taxonomy, EcoVadis, CDP, SBTi, reporting suggestions, own workforce writes, and holdings. In the public API v2 that means purchases and suppliers, wastes, business travels, hotel stays, freight transports, employees and commuting, sold products and their use, logistics, own workforce, waste water treatments, water, district heating, district cooling and process invoices, and webhook endpoints, except listing them and reading their deliveries (403 PLAN_FEATURE_NOT_INCLUDED). Review any client that calls these modules from a trial organization.
  • Manual entries: GLEC shipments, GLEC recharges, and product periods now have a source and an activity log (history and comments).
  • Fix: Wastes now apply updated LER material rules, an OpenAI-based confidence score, and factor suggestions.

October 26, 2026

API
  • ⚠️ Framework plans: Writes to framework plans are now restricted to Dcycle staff. Update any client that modifies framework plans with a non-staff account.
  • ⚠️ Legacy endpoints: The mobile testing route and the AWS_LAMBDAS authorization shortcuts have been removed. Update any client that still calls them.
  • ⚠️ Reporting suggestions: Runs are now gated by plan, and each organization has a spend cap. Review any client that triggers reporting suggestions from an organization without the required plan.
  • Wastes: POST /v2/wastes now supports reporting waste in tonnes.
  • Purchases: The free plan now includes AI description matching for up to 500 purchase lines per organization (lifetime).
  • Fix: The support ticket image proxy now serves only images from the ticket’s own HubSpot conversation.

October 25, 2026

API
  • ⚠️ Organizations: Each user now gets one free_plan organization across every creation route, not just some of them. Update any client that creates several free-plan organizations for the same user.
  • ⚠️ Supplier risk assessment (SRA): Every SRA write now requires the SRA plan. Review any client that writes SRA data from an organization without it.
  • ⚠️ Supplier risk assessment (SRA): Survey emails now use fixed subjects and are subject to a send budget.
  • Fix: Resending an SRA invite to a different target now rotates the invite token.

October 24, 2026

API
  • ⚠️ Holdings: Only an admin can now manage a holding’s subsidiaries. Update any client that manages subsidiaries with a non-admin member.
  • Fix: Only members who can manage the supplier portal can now invite suppliers.

October 23, 2026

API · MCP
  • ⚠️ Legacy endpoints: 36 legacy routes with no legitimate caller have been removed, along with the websocket and test-only routes. Update any client that still calls them.
  • ⚠️ API keys: An API key can now write only while its creator may write in that organization. Review keys whose creator has lost write access.
  • ⚠️ Roles: Member-only writes now check role permissions. CSRD sensitive data is now filtered on the server.
  • ⚠️ Free plan: Facility and subsidiary caps are now enforced on legacy routes.
  • ⚠️ x-partner header: A value that is not one plain path segment is now refused.
  • ⚠️ Purchases: In v2, the supplier country must now have purchase factors, and blank supplier names are refused. Review the request body of any client that creates purchases through v2.
  • ⚠️ MCP OAuth: OAuth clients can now redirect only to known MCP client callbacks. The user must also consent before a client receives its code. Re-check custom MCP clients that use OAuth.
  • Bulk imports: New bulk upload of intensity metric values, historical emissions, avoided emissions, and electricity and heat suppliers.
  • Logistics: The license-plate pick-list for shipments can now be searched. Vehicle ownership (subcontractor) can now be bulk-edited on shipments.
  • ESG-X: An existing standalone-tenant report can now be attached as a Dcycle project.
  • SBTi: New catalogue v2 and default dashboard.
  • Vehicles: Each fuel line’s vehicle is now suggested from its plate, booking history, a misread plate or a fuel placeholder. License plates are now compared on letters of any script.
  • Hotel stays: v2 now accepts ongoing stays.
  • Employees: v2 now normalises names and emails the same way as the app and the import.
  • Own workforce: v2 trainings, absences and employees now take dates, hours and texts the same way as the file import.
  • Freight transports: v2 now accepts what calculates. A failed section now says which one failed and why.
  • Business travels: v2 now accepts what v1 accepts. A two-character origin that is not a country code is now located.
  • Ingest jobs: A retry with the same idempotency key now answers IDEMPOTENCY_KEY_IN_PROGRESS immediately.
  • Fix: v2 bulk requests with text or numbers that previously returned 500 now return 422.
  • Fix: Invalid Excel uploads to LCA now return a 4xx instead of a 500.
  • Fix: Media URLs are now signed for API-key readers and confined to the key’s organization.
  • Fix: Key ownership checks now also cover jobs, partitions and ACV, and transient lookups are retried.
  • Fix: Attachment keys are now matched exactly, and group grants are no longer derived from client-stored URLs.
  • Fix: Heavy-upload keys that name another organization are now rejected.
  • Fix: Only the owner can now write to a personal assistant conversation, and only as the user.
  • Fix: ESG-X calls now go to the tenant that esg-x-core says serves the organization.
  • Fix: Remunerations of a new contract sent newest first no longer crash the load.

October 22, 2026

API
  • ⚠️ Legacy endpoints: Anonymous writes to legacy endpoints are now refused. Update any client that writes to them without credentials.
  • Fix: Superadmin powers now require a user session and can no longer be used through an API key.
  • Fix: Registration now decides the plan, partner and fund flag on the server instead of taking them from the request.

October 21, 2026

API
  • ⚠️ Electricity invoices: POST /v2/invoices/electricity now takes the toc of a supercharger’s electric vehicles. Review the request body of any client that sends supercharger invoices.
  • Hotel stays: New public v2 bulk create and delete.
  • Sold products: New public v2 bulk create and delete, and a public v2 bulk use.
  • Waste water treatments: New public v2 bulk measurements and v2 plants.
  • Stationary fuels: The API now lists the units each fuel is calculated in.
  • Purchases: Description-matched imports now accept supplier-less and physical-only rows. Purchases left without an emission factor now appear in the factor-review to-do.
  • Ingestions: Consumption duplicates now expose what they matched on. Files can now be found by name including the extension.
  • Files: When an original goes to the trash, a held duplicate now becomes the original.
  • Logistics: New license-plate pick-list and index for the shipments grid filter.
  • Employees: Fully remote periods are now priced on their total telework hours.
  • Own workforce: New manual forms, period management and incident custom fields.
  • Document processing: A credit note that mirrors one invoice of its reading now cancels that invoice.
  • Fix: Creating an API key is now refused when the organization has no API access.
  • Fix: An API key’s x-user-id can no longer name Dcycle staff who do not own the key.
  • Fix: Customers can no longer add Dcycle staff to their organizations.
  • Fix: Public forms now authenticate with the signed link token instead of the webhook API key.
  • Fix: A NUL character or an out-of-range number in a v2 request now returns 422 instead of a 500 or a stuck job.
  • Fix: The geocode location_type now keeps the place type returned by AWS.
  • Fix: Invoice media URLs stored with a double slash before the bucket are now signed.
  • Fix: A parent’s file shown on a subsidiary’s record is now signed.
  • Fix: The duplicate warning no longer points at an original that is in the trash.
  • Fix: A held evidence copy is now released without being processed.
  • Fix: Document splitting now judges each range by its own controls instead of the first range’s.
  • Fix: D&B suppliers are now ranked once each by EUR spend, and a looked-up supplier is never overwritten.
  • Docs: New guide on what an integrator needs to send wastes through v2, plus v2-only waste guides.

October 20, 2026

API · MCP
  • ⚠️ MCP: The remote MCP server now requires authentication and checks organization membership. Update remote clients that connect without credentials.
  • MCP: New LCA co-product and allocation tools, including set_lca_allocation. New LCA block assumption and comment tools.
  • MCP: New CDP tools for the questionnaire, answers, Leia drafts and the portal export.
  • MCP: Tool inputs now declare units, and destructive tools carry destructive annotations.
  • Business travels: New public v2 bulk create and delete, run as ingest jobs.
  • Freight transports: New public v2 bulk create and delete.
  • Own workforce: New public v2 create and delete for absences and accidents.
  • Ingest jobs: An Idempotency-Key header now identifies the route and body for 24 hours.
  • Imports: New waste water treatment bulk upload (CSV/Excel), with undo and a pre-audit overlap check.
  • Custom emission factors: Waste custom factors can now be defined per cubic metre as well as per kilogram.
  • Attachments: Wastes, waste water treatments and custom emission factors now accept a My files file_id.
  • Media: Media-bucket URLs in responses are now signed. Signed URLs renew, and links in emails stay stable.
  • EcoVadis: Leia run datapoint counts are now reported per question, not per option row.
  • SRA: Only the best-scoring answer now requires evidence. Re-importing suppliers now fills in the empty fields of existing suppliers.
  • Supplier risk: New basic questionnaire for tier-2 suppliers. Answers and evidence already given are kept when the questionnaire changes.
  • ISO 45001: New project type with a hazard register and OH&S records.
  • Rate limits: Limits now charge each caller (API key, user or IP) instead of sharing one bucket across all app users.
  • Total impacts: Validated emissions now leave out children with a 0% share and report biogenic CO2 separately.
  • Fix: A facility split may add up to less than 1 again. This relaxes the “splits must add up to 1” rule from October 19.
  • Fix: POST /v2/vehicles now accepts every fuel the vehicle class runs on.
  • Fix: v2 vehicle consumptions now accept only the unit of the vehicle’s own emission factor.
  • Fix: v2 employees now take the import’s contract dates. A remuneration dated before its contract, or an amount the column cannot hold, now returns 422 instead of 500.
  • Fix: Intensity metrics now divide each record by the emissions of its own period.
  • Fix: Redeemed GdO kWh now count as renewable in the energy mix.
  • Fix: The predefined Administrator role is now listed as scoped to every My categories dataset.
  • Fix: Deleting records with delete_source_files no longer times out on logistic requests. Deleting an evidence file now unlinks the wastes and invoices that attached it.
  • Fix: A reading waiting in the queue is no longer reported as a reading error. Old files no longer resurface as in progress when the reconciler touches them.
  • Fix: Last change and classification timeout now anchor on the upload, not file.updated_at.
  • Fix: MCP tools that always failed in production now call the backend’s real routes. set_lca_allocation input checks are tighter.
  • Fix: New subsidiaries now join auto_tree decarbonization plan perimeters, and unknown sectors are rejected.
  • Docs: The bulk flow now has a single page, and the v2 pages now describe what the code does.

October 19, 2026

API · CLI · MCP
  • ⚠️ API v2: The v2 record contract is now closed. Splits must add up to 1, and unknown fields are rejected. Update clients that send extra fields or splits that don’t sum to 1.
  • ⚠️ API v2: The single-create endpoints for vehicle consumptions are deprecated. Use POST /v2/vehicle-consumptions instead.
  • Vehicles: New POST /v2/vehicles creates vehicles in bulk, synchronously. New POST /v2/vehicles/delete deletes vehicles through an ingest job.
  • Vehicle consumptions: New public v2 bulk create and delete.
  • Own workforce: New public v2 bulk create and delete for employees (headcount). New public v2 create and delete for trainings.
  • Wastes: The v2 API now takes facilities as facility_percentages, with one code per facility rule.
  • Invoices: Credit notes are now supported on heat and electricity. The PDF reader now nets corrective invoices.
  • Integrator lookups: Every integrator lookup now answers with an API key alone. These lookups are documented outside the OpenAPI spec.
  • EcoVadis: The Data view can now be filtered to report evidences per question. Every questionnaire row now starts included in the report. Leia can now read the project’s current score estimate and explain how to raise it.
  • SRA: Inherent risk now uses tiers based on country, activity, on-site work, spend and criticality.
  • Factor review: Low-confidence purchase and waste matches now count while they wait for review.
  • Files: Opted-in presigned uploads now sign the client’s SHA-256 as the S3 checksum. Opted-in attachment uploads sign the announced SHA-256. Comment, non-conformity and custom KPI attachments now go through the upload lifecycle, and every file row is created or completed through it. Every document download now reconciles the file’s content hash with the bytes it read.
  • Reporting Suggestions: Leia’s proposals are now bounded by pages, not files.
  • Imports: The employee commuting import now accepts an end date in the current year. The combustion facility-consumption template now has a CUPS column.
  • Document Processing: The document splitter is now gated for wastes and transport too. A processing task can now read several documents at once.
  • Fix: The public v2 spec now requires only the fields that list responses send, and declares the bulk headers.
  • Fix: Editing a split waste now rewrites every part and its split.
  • Fix: Bulk editing logistics legs no longer returns 500 when the post-commit refresh fails.
  • Fix: Recharge invoices created from file readings now keep the reviewed amount and currency.
  • Fix: CSV import workers now refuse another tenant’s custom emission factor. Bulk-split now reports why an invoice failed.
  • Fix: Employee responses submitted anonymously no longer reveal who created the employee.
  • Fix: Custom emission factor databases with no factors can now be deleted.
  • Fix: The CDP 2026 catalogue was rebuilt from the portal and the 2026 guidance.
  • Fix: The SRA inherent risk is now recomputed after suppliers are imported from Purchases.
  • Fix: Document Processing now matches vehicle-fuel plates printed with a cost-centre suffix.
  • Fix: Purchase OCR now assigns a physical unit only to goods, and resolves kgs, lts and mtr.
  • Fix: The EcoVadis score estimate now follows the Scoring Principles V1.7 literally.
  • Fix: A PDF that PageIndex Flash refuses is now reported as a problem with the file, not as a provider error.
  • Docs: The API Get Started guide now leads to v2, and the v1 guides carry a deprecation warning.
  • Docs: The MCP example now works as written, unpublished pip installs are removed, and the CLI is referred to as dcy throughout.

October 18, 2026

API
  • ⚠️ API v2: POST /v2/logistics/requests is now the single route for creating one or many legs. Update clients that call /v2/logistics/requests/bulk.

October 16, 2026

API
  • ⚠️ API v1: POST /v1/purchases and DELETE /v1/purchases/{purchase_id} are deprecated. Use POST /v2/purchases and POST /v2/purchases/delete instead.
  • Vehicles: PATCH on a vehicle can now move it to a descendant organization.
  • Logistics: You can now edit shipments in bulk, including electric transport operation categories (TOCs).
  • EcoVadis: New score estimate engine with coverage, evidence documents and coverage reading.
  • CDP: New CDP project type with the 2026 questionnaire catalogue and a default dashboard. Leia can now answer the CDP questionnaire.
  • SRA: Leia can now propose remediation actions for a supplier.
  • Employee: Commute kilometers are now editable. A new history feed shows Leia’s distance estimates.
  • Reporting Suggestions: Leia extraction and writer now run on the Standard tier. CSRD Leia is open to csrd:general, and S1 and closed documents follow the member’s role.
  • Reporting Suggestions: A re-analysis requested by a person now reopens an answered datapoint.
  • Files: A missing content_hash is now repaired from the stored object, and Leia reads those files. One upload lifecycle now settles the content_hash of every RAW file.
  • LCA: The matching outcome now counts closest approximations separately.
  • Document Processing: The document splitter is now skipped when classification saw only one document.
  • Fix: Editing a logistics leg without distance or load now returns 400 instead of 500.
  • Fix: POST /v2/logistics/requests now rejects records that have no distance or no addresses.
  • Fix: Transport geocoding now treats any trailing country other than ES or PT as foreign.
  • Fix: A use-of-product quantity of 0 is now treated as “block does not apply” in imports.
  • Fix: content_hash is now stored on every path that holds the file bytes. Portal and heavy uploads keep the client hash, and invoice and waste uploads keep their hash.
  • Fix: The duplicate file check now separates identity from eligibility.
  • Fix: EcoVadis files without a content_hash keep their full reading and are no longer read in a loop.
  • Fix: Reporting suggestions now skip a source without a content_hash instead of failing the run.
  • Fix: Document evidence now reads ligatures correctly when Word’s ToUnicode map breaks them.
  • Docs: The v2 logistics bulk page now documents the 409, the distance error, the rare per-leg error codes, field reading, the business key and auth answers.

October 15, 2026

API · CLI · MCP
  • ⚠️ API v2: Invoices now take facilities only as facility_percentages. POST /v1/invoices and DELETE /v1/invoices are deprecated. Use the /v2/invoices endpoints instead.
  • API v2: POST /v2/purchases and POST /v2/purchases/delete now run as ingest jobs.
  • API v2: POST /v2/logistics/requests/bulk creates logistics requests as an ingest job.
  • EcoVadis: New catalogue of 21 criteria, with manual criteria configurable per project. Scoring settings are now per project, and you can download the documents that back an answer. Document checks are more precise: automatic full read, month and year dates, OCR scans, similar entity names and GHG verifications.
  • EcoVadis: New EcoVadis CLI commands and read-only MCP tools.
  • Custom KPI: New number_breakdown KPI type, which holds an amount per category.
  • Custom KPI: New custom KPI write commands in the CLI and MCP. Goals and actions now support owners.
  • Action Plan: Action plan alerts now use server-side goal rules. A daily job sends notifications and an email digest.
  • Files: You can now confirm the set of services of a multi-service file in one call.
  • Purchases: Purchase invoices read by OCR now include physical quantities and the matching method.
  • Reports: GHG annexes now disclose wastes.
  • Projects: Projects now expose the organization name that the report covers print.
  • Activity Feed: Wastes, hotel stays and historical emissions are now versioned for the activity feed.
  • Attachments: Manual waste and invoice attachments now keep their original file name.
  • LCA: Leia now always links a resource to its closest ecoinvent market. It also warns when a material is linked to a processing step only, and when small electronics have no circuit boards. Transport is no longer treated as a process.
  • MCP: New tools to edit LCA blocks, quantities and the functional unit.
  • Logistics: Own electric trucks now get their supercharger’s electricity in GLEC calculations.
  • Document Processing: Extraction now uses gpt-6-luna.
  • PPWR: The single-use plastic grouped ban is now a warning until 2028, and critical from then.
  • Fix: A concurrent ingest job request with the same idempotency key now returns 409 with IDEMPOTENCY_KEY_IN_PROGRESS.
  • Fix: GET /v1/invoices/{id} now returns the co2e written by the calculation.
  • Fix: The public OpenAPI spec now authenticates with its declared ApiKey and Organization schemes.
  • Fix: A reading still in flight is now reported as processing, not as a reading error.
  • Fix: Waste and wastewater treatment (WWT) attachments sent as data URIs are now uploaded and keep their original name.
  • Fix: Regenerating a report now keeps blocks that repeat a template slot.
  • Fix: Facility-consumption suppliers in imports are now scoped to the organization family.
  • Fix: The facility matcher now requires a street-name match in its proximity layer. Candidates are ranked by shared street before the cap is applied.
  • Fix: Long grouped invoices are now read correctly, including windows, bundle numbers and the summary item.
  • Fix: The MCP list_units tool now sends the organization header.
  • Fix: MCP LCA block reads now flag hidden materials. Edits now refuse incomplete rows.
  • Docs: The API v2 invoice docs now include lookups, definitions and links.
  • Docs: The API v2 docs now list the standard resources and auth errors in one place. The wastes page is now the template for other resources.
  • Docs: New List Purchase Suppliers page. The filter_by operators of List Purchases are corrected.

October 14, 2026

API
  • Purchases: Purchases are now matched to emission factors by description, for both spend-based and physical factors. OCR-reviewed purchases are matched by description by default.
  • Purchases: Reviewed purchase invoices can now be applied as purchases. The supplier already confirmed for an issuer tax id is now proposed.
  • Document Processing: New invoice::purchase category reads supplier purchase invoices from PDF. The PDF classifier now offers it.
  • Document Processing: New document::risk_assessment category, plus agreement and initiative endorsement categories. They support manual-only reading and review.
  • Document Processing: PDFs that glue several bills together are now read one bill at a time.
  • Document Processing: EcoVadis evidence categories are now classified for every organization (prompt v10, Decisions contract v5).
  • SRA: The questionnaire can now be sent to many suppliers in the background.
  • SRA: Suppliers can now be homologated as approved, approved with conditions or rejected, with a review date.
  • Custom KPI: New duplicate campaign endpoint. Campaigns now have a launched (draft) state.
  • Voluntary Standard: New VSME project type with its catalogue and vsme.report template. Leia reporting suggestions are now backed by a VSME RAG corpus.
  • Reporting Suggestions: Leia now supports EINF (Ley 11/2018) projects, with required elements and a Sol writer.
  • Reports: New opt-in GHG annex that lists the value of each emission factor.
  • LCA: A4 export rows now include route names. B6/B7 dashboard contributors are now broken down per country.
  • Emission Factors: Goods and employee emission factor rows now return shipment and person ids.
  • FLAG: The crop-inputs list now returns per-input CO2e for scope 1 and scope 3.
  • Report Builder / Goals: You can now choose how measures and KPI answers combine where they are used.
  • Transport: electric and refrigerated are now optional in the shipments bulk upload and the create API.
  • Fix: Ingest job chunks interrupted by a worker shutdown are now retried. Fuels ingest jobs now process chunks of 50.
  • Fix: Purchases now apply a fresh physical match even when the “embedded” cache is stale. Stored description-match decisions, factor guardrails and match follow-ups are also fixed.
  • Fix: A file keeps its guided category when a held duplicate is resolved.
  • Fix: Text extracted from PDFs no longer contains line-break hyphen artifacts.

October 13, 2026

API
  • API v2: New POST /v2/invoices/electricity, POST /v2/invoices/fuels, POST /v2/invoices/refrigerants and POST /v2/invoices/water create invoices by category.
  • API v2: New POST /v2/invoices/delete deletes invoices through an ingest job. New POST /v2/invoices/process processes invoices.
  • API v2: POST /v2/facilities has been trimmed, and unit is now surfaced as a reference to units.
  • Files: A file can now hold several classifications, each with its own reading. A file with several services now waits for confirmation as a whole.
  • File Reading: Each reading is now linked to its classification. The link is expand-only.
  • Document Processing: A PDF that bills several utilities is now classified as each of them. Utility extraction is now scoped to the requested service, and a processing command confirms the services it reads.
  • Document Processing: New document::training_record, kpi_report, ghg_inventory and cdp_response categories. They support manual-only reading and review.
  • Document Processing: Annual accounts and annual reports are now classified for every organization.
  • Document Processing: New billing-period prompt rules for invoices and stationary combustion.
  • File Reading: Confirmed suppliers are now remembered per organization (org_history). A manual supplier choice is also remembered on readings created before the matcher existed, and the reading’s country is kept on review edits.
  • Invoices: Placeholder invoice numbers no longer group duplicates. Overlap warnings now carry a group id.
  • Emission Factors: You can now filter the purchases grid by unit.
  • Imports: Import templates now include per-column help and an example.
  • Imports: SRA supplier imports now report skipped rows. A single SRA supplier cell now accepts several emails.
  • EcoVadis: The Leia question picker is now grouped by question. It is ordered by theme, then Policy, Action and Reporting.
  • EcoVadis: Analyze and classify now filter 1. Files also gain generic duplicate inheritance and post-commit listeners.
  • Elastic Data: You can now hand-edit a dataset’s saved Leia prompt.
  • Management Systems: The project action plan now includes the goals’ KPI history, pending readings and actions.
  • SRA: Evidence is now required only for answers that score.
  • Fix: Editing a logistics shipment no longer converts its stored kg load a second time.
  • Fix: Combustion invoices read from a PDF are now marked with source pdf.
  • Fix: The document detail for an ingestion now names the assigned supplier.
  • Fix: Numeric CNAE codes are no longer flagged as a text-field mismatch on import.
  • Fix: Multi-database waste LER and R/D treatment codes are now read from the kernel catalogues in pivots and widgets. The pivot and widget caches have been invalidated so results refresh.
  • Fix: Employee recalculation now also creates missing telecommute configs.

October 9, 2026

API
  • API v2: New POST /v2/facilities creates facilities in bulk and responds synchronously. New POST /v2/facilities/delete deletes facilities through an ingest job.
  • Reporting Suggestions: You can now GET a run’s targets, the datapoints the run was asked to answer.
  • EcoVadis: Only one questionnaire can be active at a time per project. Projects go back to the standard questionnaire, and Leia now reads the standard questionnaire.
  • EcoVadis: Document-check evidence is now located, and you can exclude documents from the analysis.
  • EcoVadis: A run over new documents now keeps the earlier proposals that the new run did not find.
  • EcoVadis: The question catalogue now includes a source provenance column.
  • Purchases: The purchase type now follows the applied emission factor.
  • Transport: Bicycle is now available as a mode for goods sent and received.
  • File Reading: A confirmed electricity invoice’s CUPS is now listed on its facility.
  • Historical Emissions: wastes_transport entries are now accepted and reported.
  • Logistics: TEU/FEU load units are now calculated at the GLEC average load in the API. Unknown load units are now logged.
  • Security: The supplier catalogue CSVs and the legacy back-office supplier routes are now restricted to staff.
  • Fix: Duplicate invoice warnings now require the same consumption, not just the same invoice number.
  • Fix: Evidence uploads no longer appear in the active ingestions feed.
  • Fix: EcoVadis document checks computed when a Leia run starts are now persisted.
  • Fix: Logistics TOC filters now apply to the resolved toc_id. Page ids are now selected before joins, and the TOC pick-list is built from the TOCs in use.
  • Fix: Reporting Suggestions now handle not-found cases more robustly.
  • Fix: Creating, editing or deleting a historical emission now flushes the dashboard caches.

October 7, 2026

API · Documentation
  • Purchases: Purchases can now be negative, to record returns. A net guard is applied to the result. Imports also read a typographic minus (−) as a negative amount and refresh the negative_purchases flag when a file is validated.
  • Flag: Synthetic fertilizers of the complex and other types are now supported, using a user-provided %N.
  • Emission Factors: Coincidence options are now served for every category.
  • EcoVadis: output_data is reworked into a deduplicated question catalogue, with evidence value types.
  • EcoVadis: When a document entity does not match, the error now names the rated companies.
  • Reporting Suggestions: Review conflicts now return explicit conflict codes. Each suggestion also reports whether it can be restored or answered.
  • Document Processing: Audit reports can now be read and reviewed through the sot_audit_report document type (manual-only).
  • API v2: Webhook deliveries are now easier to debug. Webhooks are sent only for jobs created with an API key. Wastes now accept kilograms only.
  • Ingest Jobs: The app job source is now accepted as a valid value.
  • SoT v2: Custom output data is now reported by default.
  • Fix: CSRD reports now include custom CSRD output data that is flagged to report.
  • Fix: Datasets queries now cover any project the organization may open.
  • Fix: Datasets that hold stationary and mobile fuel in the same table no longer return empty results.
  • Fix: Reporting v3 no longer prorates the ENERGY column twice for invoices that cross a year boundary.
  • Fix: Document processing now leaves a stationary fuel for review when it contradicts the printed product.
  • Docs: The units reference now describes what each type returns, the real error responses and exact-name lookups.
  • Docs: The API v2 documentation now has working error examples and shows where to find every ID. Problem details are now included in the public OpenAPI specification.

October 5, 2026

API
  • LCA: The by-filters resources actions now verify a filter_hash before they run.
  • Employees: Surveys can now be resent by filters, using filter_hash to identify the filtered selection.

October 2, 2026

API
  • EU Taxonomy: New perimeter API to manage projects and activity decisions.
  • Logistics: Recharges and requests now support editing of all their fields.
  • Auth: A role’s dataset scope now refuses the sidecar of a native category.
  • Fix: Consecutive invoices that share a boundary date are no longer flagged as overlapping during ingestion.
  • Fix: Deleted invoices are no longer re-created from their original file.
  • Fix: Corrected two impact units for ecoinvent 3.11 in LCA reports. The report builder now also states what the LCA holds.

October 1, 2026

API
  • Emission Factors: errors_total on POST /v1/emission-factors/{category}/config-get-all now counts only the rows with no emission factor (config_source unmatched). A record that has a factor but is itself in error (record_status error) is no longer counted; its record_status is still returned.

September 2026

September 30, 2026

API · Documentation
  • LCA: Automatic ecoinvent matching now proposes and applies market activities only.
  • Emission Factors: POST /v1/emission-factors/{category}/config-get-all now lists records for which no emission factor could be found, with config_source unmatched, so a database can be chosen for them. Each row also returns record_status (the record’s own status, e.g. error), and the page returns errors_total, the number of rows in error across all pages.
  • Emission Factors: GET /v1/emission-factors/{category}/candidate-matches now returns activity_categories on each candidate, the requested categories the match applies to. PATCH /v1/emission-factors/{category}/config-update rejects a match from a different category with 422 CONCEPT_NOT_IN_CATEGORY. It also rejects a custom record that only has per-gas factors (CO2, CH4 or N2O) for a category that can only use CO2e, with 422 CUSTOM_RECORD_REQUIRES_CO2E.
  • ⚠️ Emission Factors: PATCH /v1/emission-factors/employees/config-update (and config-update-by-filters) now returns processing_job_id instead of saved_ef_count / skipped_ef_count, like hotel_stays.
  • Fix: PATCH /v1/emission-factors/employees/config-update no longer blocks the whole batch when the chosen database has no match for one commuting period. Uncovered periods keep their current factor and are listed in the processing job’s failed_records (COMMUTE_MATCH_NOT_FOUND, with employee_name). 400 EF_SOURCE_NO_EMPLOYEE_MATCHES is returned only when no selected period is covered.
  • Fix: Saving an invoice no longer resets the scope 3 databases chosen on the emission factors screen.
  • Fix: Changing an invoice’s dates no longer leaves it listed on the emission factors screen under years it no longer covers.
  • Documentation: Clarified how POST /v2/wastes finds the emission factor for a waste record.

September 29, 2026

API · Documentation
  • ⚠️ Wastes: POST /v2/wastes now accepts custom_ef_record_id and no longer accepts custom_emission_factor_id. Integrations that send custom_emission_factor_id must switch to custom_ef_record_id.
  • Wastes: POST /v2/wastes no longer applies admission control. A single POST now handles one or many waste records.
  • LCA: POST create on the v1 LCA endpoint now returns the ecoinvent matching job it started.
  • LCA: Recalibrated the ecoinvent match confidence thresholds. A match is now high at 0.60 and above and medium at 0.50 and above.
  • Reports: A report’s language can now be changed after creation.
  • Reporting suggestions: Word (.docx) sources are now accepted. Each one is converted once to a frozen PDF.
  • Invoices: The invoice list and detail responses now return the uploaded file name.
  • Fix: Invoices with overlapping periods but a different fuel are no longer flagged as duplicates during ingestion.
  • Fix: Emission factor responses now show the config source for records without conversion_steps.
  • Documentation: Documented the custom emission factors API (BETA) and linked it from the wastes documentation.

September 28, 2026

CLI · MCP · Documentation
  • CLI: Added dcy elastic (also dcy categories) to manage My categories from the terminal: create, edit and delete datasets, fields and records. record create takes one --data object or a --file JSON array and reports each row’s result; record list supports --filter and --sort. Deletes ask for confirmation unless --yes is passed. Requires dcy auth login; API keys are rejected. Available from dcy v0.0.127.
  • MCP: Added write tools for elastic data: create_dataset, update_dataset, delete_dataset, add_dataset_fields, update_dataset_field, delete_dataset_field, create_dataset_records (up to 100 rows per call, with per-row results), update_dataset_record and delete_dataset_record. Delete tools are annotated destructiveHint: true. They require the OAuth (remote) connection.
  • Documentation: Added the My Categories CLI page and documented the new elastic data write tools.

September 26, 2026

API
  • ⚠️ Authentication: Requests that set the x-organization-id header are now rejected unless the caller is allowed to act as that organization. Previously, the header was not fully checked against the caller’s permissions. Integrations that switch between organizations must use credentials authorized for each target organization.
  • ⚠️ Reports: Every route that takes a {report_id} is now scoped to the report’s project perimeter. Requests for reports outside the caller’s perimeter, including reports belonging to other organizations, now return 404.
  • ⚠️ Projects: Project configuration changes that manage the perimeter now require permission to manage that perimeter. Organization links are now limited to organizations within the caller’s group. Requests that reference organizations outside the group return 404.
  • ⚠️ Reports: Dashboard comments and widget share projects are now authorized against the caller’s perimeter. Requests for resources outside it are rejected.
  • Security: Entity IDs passed in paths and request bodies are now validated against the caller’s organization across legacy routes. Requests that reference another organization’s resources are now rejected. Affected resources include:
    • Emission factor overrides and emission factors
    • Logistic hubs, transport, and purchases
    • LCAs, including LCA IDs sent in request bodies
    • Funds and ESG
    • Validations, events, and data groups
    • Destination IDs in request bodies
    • Organization-ID routes, the sanitise-all operation, and /emails
  • Support: Sending an unrecognized x-partner header now returns 400 with error code UNKNOWN_PARTNER. Previously, it returned 500.
  • Transport: Bulk air and sea rows without coordinates now return a ZERO_RESULTS geocoding status instead of failing with an internal error.
  • Transport: Maritime distances from the Canary and Balearic Islands to other countries are now calculated over sea routes. Distances and emissions for these shipments may change as a result.
  • Facilities: The group view facility list now correctly counts the categories without data for each facility.
  • FLAG: Updating land management data now recalculates emissions for the facilities and vehicles linked to the same family.
  • Holdings: Deleting a single employee now clears the contract limit flag when the organization drops back under its limit.

September 25, 2026

API
  • Datasets: The Social dataset now includes an Edad al salir (age at departure) field for departed employees.
  • Datasets: The Social dataset now exposes the legacy pivot table’s social measures alongside its own, so reports built on the legacy pivot can be reproduced from the dataset.
  • CSRD Reporting Suggestions: Generated suggestions now state data absences accurately and use tenses that match the reporting period. A new faithfulness check validates each suggestion against its source documents before it is returned.
  • CSRD Reporting Suggestions: Document checks are more reliable. Runs now handle checks that finish late, documents that reference group entities, storage timeouts, failed source-reference lookups (retried), and corrupt profiles without failing the whole run.
  • Transport: Shipment addresses extracted by OCR now fall back to an alternative geocoding strategy when the primary lookup fails. This reduces unresolved origins and destinations.
  • Datasets: CH4 and N2O values are now read in actual tonnes. CO2 is now split into separate fossil and biogenic measures. Reported CH4 and N2O figures may change as a result.
  • Invoices: Legacy water invoices are now priced in m³, regardless of the unit they were originally recorded in. This corrects emissions for water invoices stored in other units.
  • Business Travel: Trips reviewed through OCR are now linked to their source file. Deleting the file now also removes its associated trips.

September 24, 2026

API
  • ⚠️ Projects: Creating an Ecovadis project now requires a purchased Ecovadis framework plan. Organizations without the plan can no longer create Ecovadis projects.
  • Projects: Ecovadis is now a full reporting framework, handled the same way as the other supported frameworks.
  • Projects: New endpoints let you upload and delete a custom project image.
  • Management Systems: Added an ISO 14001 legal register per facility (phase 1) for tracking applicable environmental legal requirements.
  • Data Collection: Data collection tasks can now be shared across projects. This release also adds EINF eligibility and category coverage information for tasks.
  • OCR: Transport invoices can now be read with OCR. OCR for business travel and transport is now fully active.
  • Business Travel: Flight and train tickets can now be read with OCR to create business travel entries automatically.
  • CSRD Reporting Suggestions: Each suggestion now includes the datapoint requirement it answers.
  • CSRD Reporting Suggestions: Suggestions can now flag documents that describe a different entity or an older reporting period.
  • CSRD Reporting Suggestions: Results are now delivered one by one as each is produced, rather than all at once when the run finishes.
  • CSRD Reporting Suggestions: The run input now specifies the reporting framework, and each framework has its own profile.
  • CSRD Reporting Suggestions: Suggestions are now written in report-ready language and generated from a more compact context.
  • Datasets: The invoices and emissions datasets now support a scope 2 location-based reading alongside the market-based one.
  • Datasets: Activity tons_co2e is now split into the activity’s own emissions and its scope 3 category 3 (upstream fuel- and energy-related) portion.
  • Datasets: Dataset visualizations are now scoped to their project, matching the behavior of the legacy pivot table.
  • Waste: The default emission factor chain is now DEFRA → OCCC, with no fallback to Legacy factors. Gaps in OCCC are filled using equivalent factors from the matching hazard class. Some waste emissions may change as a result.
  • Waste: Waste items extracted by OCR without a LER code are now queued for matching based on their description.
  • LCA: Every piece_impacts row now records the full calculation (all multiplication factors) behind its impact value, for complete traceability.
  • Waste: The total count returned by GET /v1/wastes is now correctly scoped to project_id.
  • Invoices: The reviewed end date on fuel-delivery invoices is now preserved and no longer overwritten.

September 23, 2026

API
  • CSRD Reporting Suggestions: Citation highlight rectangles are now resolved and returned for reporting suggestions, enabling precise visual anchoring of extracted content to source documents.
  • CSRD Reporting Suggestions: Suggestion runs can now be paused and resumed via POST /v1/csrd/reporting-suggestions/runs/{id}/pause and POST /v1/csrd/reporting-suggestions/runs/{id}/resume.
  • CSRD Reporting Suggestions: Stopped suggestion runs can now be deleted via DELETE /v1/csrd/reporting-suggestions/runs/{id}.
  • CSRD: Citation validation now rejects sign flips and invented middle sections in citations with empty-edged ranges, ensuring citation integrity.
  • Emission Factors: the goods transport categories (transport_distribution_downstream, transport_distribution_upstream) can now be filtered by description. GET /v1/emission-factors/{category}/unique-values accepts field=description (it returned 400), and POST /v1/emission-factors/{category}/config-get-all accepts a description list in the body (it was ignored). The description is the route name.

September 22, 2026

API
  • Transport: Weight units now support pounds (lbs) in addition to kilograms and metric tons. The weight_unit field accepts lb and automatically converts to metric for emissions calculations.
  • Waste: POST /v1/pivot-table now splits waste CO2e emissions into separate measures for treatment emissions and collection transport emissions, providing granular breakdown of waste-related impacts.
  • Waste: Low-confidence emission factor matches are now applied to waste calculations while maintaining their flagged-for-review status, preventing calculation gaps while preserving data quality signals.
  • Waste: Column filters for ler_code and rd_code fields now correctly filter results in GET /v1/wastes responses.
  • Vehicle Fleet: Vehicle data entry now uses native reading inputs instead of the PDF importer flow, streamlining data collection.
  • Data Imports: Header column reranking during file imports now uses the Luna reranking model for improved accuracy.
  • Projects: Bulk-adding organizations to a project via POST /v1/projects/{id}/organizations/bulk-add now deduplicates organization_ids before processing.

September 21, 2026

API
  • Emission Factors: the coincidence returned by POST /v1/emission-factors/{category}/config-get-all is now the matched concept’s own stored label for the transport, transport_generation and use_of_product categories. It used to be recomposed from the source’s table, and a column name shared by both sides of that join could shadow the source’s wording with one of Dcycle’s internal keys — MITECO vehicles read maritime - Gasóleo instead of Marítimo - Gasóleo, and rows priced off DEFRA fuels read gas_oil instead of the DEFRA fuel name. Rows whose concept had no recomposition rule were dropped from the response altogether; they are now returned. Other categories still recompose the string and are unchanged for now.

September 18, 2026

API
  • Data Collection Campaigns: Task status is now automatically reconciled in both directions. When a task is marked complete in the data collection system, the underlying entity (e.g., invoice, consumption) is marked complete; conversely, when data is entered directly, the task status updates to reflect completion.
  • Data Collection Campaigns: Company strategy policy data can now be collected as a data collection task. Use with POST /v1/data-collection/campaigns and related campaign endpoints.
  • CSRD Goals & Actions: Goals and actions can now be linked to Impact, Risk & Opportunity (IRO) records, ESRS data points, and KPIs. Supported in POST /v1/csrd/goals, PUT /v1/csrd/goals/{id}, POST /v1/csrd/actions, and PUT /v1/csrd/actions/{id}.
  • Purchases: GET /v1/emission-factors/purchases/unique-values now supports field=product_catalogue to retrieve the list of available product catalogue values for purchase emission factor filtering.
  • Wastes: GET /v1/emission-factors/wastes/config-get-all now returns LER (European Waste Catalogue) attributes and common waste type names in the response dataset.
  • Facilities: GET /v1/facilities/{id} now includes the organization_name field in single-facility responses.
  • Pivot Tables: POST /v1/pivot-table now validates that requested measure_id values are valid for the specified category and returns an error for invalid combinations.
  • Transport Routes: Route editing operations now properly preserve associated calculation configurations instead of orphaning them.
  • Employees: Employee commute emissions estimation now correctly sums only the co2e value instead of incorrectly aggregating all greenhouse gas rows.
  • Transport Shipments: Shipment emissions aggregation now correctly sums only the co2e value instead of incorrectly aggregating all greenhouse gas rows.

September 16, 2026

API
  • Water Invoices: New supplier field now available on water invoice objects. Supported in GET /v1/invoices, POST /v1/invoices, PUT /v1/invoices/{id}, and POST /v1/invoices/water/filter.
  • District Heating Invoices: New district_heat_and_steam invoice type now supported, mapped to the heat_steam_and_cooling emissions category. Use with POST /v1/invoices and related invoice endpoints.
  • Use-of-Product Consumptions: Bulk import now available via POST /v1/imports/use-of-product-consumptions. Upload CSV files containing consumption data for B1 product-use calculations.
  • Facilities: GET /v1/facilities now supports server-side column sorting. Pass sort_by and sort_order query parameters to control result ordering.
  • Data Collection Campaigns: Campaign tasks can now override assignees at the campaign level. POST /v1/data-collection/campaigns and PUT /v1/data-collection/campaigns/{id} accept per-campaign assignee configuration, and tasks can be linked to specific projects.

September 15, 2026

API
  • Wastes: the wastes emission-factor listing can now be narrowed by LER code. Pass low_code (a list of exact European Waste Catalogue codes) in the body of POST /v1/emission-factors/wastes/config-get-all, and read the available codes for a year via GET /v1/emission-factors/wastes/unique-values?field=low_code. Applies to both the wastes and waste_transport categories.

September 14, 2026

API · MCP
  • Water Invoices: New fields water_meter_number and water_supply_point now available on water invoice objects (mirroring the electricity CUPS pattern). Supported in GET /v1/invoices, POST /v1/invoices, PUT /v1/invoices/{id}, import/export flows, OCR extraction, and completeness tracking. Both fields also available as filters via POST /v1/invoices/water/filter.
  • LCA Dashboard: GET /v1/lca/{id}/dashboard and GET /v1/lca/{id}/block-impacts now return impacts_pending boolean indicating when an async recalculation is in progress (e.g., for product-use B6/B7 or end-of-life C4 deferred jobs). When true, displayed impacts may change on their own as the background job completes.
  • Supplier Risk Assessment: GET /v1/supplier-risk/dashboard returns expanded v2 dataset including risk type distribution, heat map data (risk level × assessment status grid), collection funnel metrics (pending/in-progress/completed counts), and 12-month score history with month-over-month deltas.
  • Waste Emissions: Waste transport calculations using DEFRA HGV freight factors now include per-gas breakdown (co2, ch4, n2o, co2e) in emission detail responses, matching the granularity of ecoinvent factors. Previously only co2e aggregate was returned.
  • Electricity Invoice Steps: Invoice calculation steps for electricity now display DEFRA emission factor matches in the “Coincidencia” column for both production (generation/WTT) and transmission & distribution factors. Previously these matched correctly but rendered as ”-” in the steps table.
  • Copilot MCP: Dcycle MCP server’s view-contract registry adds three new queryable views — business_travels (with computed CO2e per trip), purchase_suppliers (aggregated with emissions sum and purchase count), and employees (with current commute emissions). Query via the dcycle_query_data MCP tool with view_name parameter.

September 11, 2026

API
  • Leia Reports: GET /v1/decarbonization-plans/{id}/report-artifacts now includes artifact_version field on each report artifact, enabling clients to track version history and detect when reports are regenerated.
Documentation

September 10, 2026

API
  • Supplier Risk Assessment: New supplier document management endpoints — POST /v1/suppliers/{id}/documents uploads attachments with metadata, GET /v1/suppliers/{id}/documents lists them, and DELETE /v1/suppliers/{supplier_id}/documents/{id} removes them (row-authorized per organization). Supplier contacts can now be managed via PUT /v1/suppliers/{id} with the contacts field (array of contact objects with name, email, phone, role).
  • Data Collection Schedules: POST /v1/data-collections and PUT /v1/data-collections/{id} now accept period_start, period_end (the reporting period covered by the collection), send_time (scheduled delivery timestamp), and recurrence (cron expression for recurring collections). GET /v1/data-collections/{id} returns these fields.
  • CSRD Measurement Campaigns: ESRS projects can now be assigned a Data Management Activity (measurement campaign) via PUT /v1/sot-v2/projects/{id} with the measurement_campaign_id field. The assigned campaign context is included in GET /v1/sot-v2/projects/{id}/output-data responses for BP1-04 and related datapoints.
  • Custom Emission Factors: Custom emission factors defined at a parent organization are now automatically available to all descendant organizations in the holding family. A child organization’s emission-factor listing includes the parent’s custom factors; previously they were scoped to the creating organization only.
  • Project Assistant (behind ENABLE_PROJECT_ASSISTANT flag): New POST /v1/projects/{id}/assistant/chat endpoint provides read-only Q&A over project data (invoices, emissions, suppliers) — responses include citations and confidence scores. Conversations are ephemeral (not persisted).
  • Purchases: Formula columns and dataset queries (POST /v1/datasets/purchases/query) can now reference physical_quantity (the purchased amount in physical units, separate from monetary value) — previously this field was present but unavailable to formula expressions.
  • Employee Surveys: POST /v1/employees/{id}/survey-responses now allows transport mode fields (vehicle_type, distance_km, etc.) to be omitted when the employee reports no commute — previously these were incorrectly marked as required.
  • Formula Columns: Nested conditional expressions (if(condition, if(...), ...)) are now validated at schema creation time and preserve correct calculation semantics when evaluating measure references inside branches — previously, deeply nested conditionals could produce incorrect aggregations.
  • ⚠️ Security: Removed unauthenticated operator endpoints from the public API — GET /health/data-status, POST /notifications/cache/clear, and several debug routes now require authentication. Clients calling these endpoints without credentials will receive 401 Unauthorized.
  • Documentation: Added a GET /v2/units reference page — it documents the type filter, the whole (unpaginated) catalog and how to resolve a currency id. Invoices and purchases docs now link to /v2/units instead of the legacy /api/v1/units endpoint.

September 9, 2026

API
  • Datasets & Elastic Data: Formula columns extended with additional capabilities across multiple updates:
    • Formula columns (type: "formula" in elastic data fields) may now evaluate to text values in addition to numeric results — the return type is inferred from the expression tree. Text formulas support if conditions over dimension values and string concatenation.
    • Sidecar formula columns (on related datasets or the category’s native sidecar) can reference the primary category’s own columns and read text values behind foreign-key references (e.g., [Unit.name] resolves to the unit’s display name, not its ID).
    • Row-mode formula metrics in POST /v1/datasets/{id}/query now support per-unit measure references when aggregating quantity fields.
    • Visualization schema responses (GET /v1/datasets/{id}/schema) now include formula evaluation errors, condition branch values, and defaults from the visualization’s sidecar configuration.
  • Purchases: the purchases emission-factor listing can be narrowed by its custom columns. The filter travels in the custom_values map in the body of POST /v1/emission-factors/purchases/config-get-all, not as one query parameter per column.
  • Pivot Tables: Added telecommuting_days and vehicle_occupancy measures to the pivot table measure registry — available via POST /v1/pivot-table/query for employee commute and vehicle analysis.
  • CSRD Reporting: Added POST /v1/sot-v2/projects/{id}/output-data/consolidate endpoint, which uses AI to consolidate multiple group member answers for a single CSRD datapoint into one recommended value with supporting rationale. Returns consolidated text, confidence score, and source attribution.
  • Copilot: Enhanced query planning with view contract support — POST /v1/copilot/plan now accepts a view_contract schema describing available columns, types, and filterable dimensions per page (invoices, waste), and generates filters scoped to that schema. Confidence-level filtering added for invoices and waste queries (e.g., confidence >= 0.8).
  • LCA: Fixed resource factor download to include ACV-derived (lifecycle assessment) resource factors that were previously omitted from the export.

September 9, 2026

API
  • ⚠️ Waste transport emissions promoted to dedicated waste_transport pipeline category with its own validated view, processing job entity, and DEFRA→ecoinvent sourcing — previously calculated inline with waste disposal. Existing waste invoices remain in the wastes category; the split applies to new calculations and imports.
  • ⚠️ GdO redemptions now enforce a 100%-of-consumption cap — POST and PATCH validate that the sum of all redemptions for a (CUPS, year) pair cannot exceed the actual electricity consumption in kWh. Returns 422 with REDEMPTION_EXCEEDS_CONSUMPTION when the cap is violated.
  • Energy: Added GET /v1/energy/summary endpoint returning kWh + original-unit totals (stationary, mobile, electricity, self-consumption) grouped by year/category/fuel/unit for a given organization and date range. The end_date filter is exclusive, and multi-year consumptions are attributed by year. See Energy.
  • GdO: Added GET /v1/gdo-redemptions/pending-summary endpoint returning all (CUPS, year) pairs across the organization family with electricity consumption not yet fully covered by GdO redemptions, ordered by remaining kWh. Powers the “Available to redeem” dialog.
  • Datasets: Added formula metrics — user-written expression trees over dataset numeric fields, evaluated per-row and aggregated (mode: "row") or evaluated over already-aggregated measures (mode: "aggregate"). Supports arithmetic (+, -, *, /, **), functions (round, min, max), comparisons (=, !=, <, <=, >, >=), and if(condition, then, else) with text-literal or numeric tests. Row-mode formulas may reference dimensions in conditions (e.g., if([Unit] = "G", weight / 1000, weight)); aggregate-mode formulas may nest row-mode formulas of the same request. Constants are bound as parameters; division guards against zero; NULL propagates. Requested via formula_metrics array on POST /v1/datasets/{id}/query. See Datasets Query.
  • Elastic Data: Added formula column type — a virtual column computed from the row’s other columns on every read (grid page, sort, filter, export, visualizations). The options.expr field stores a FormulaNode tree (same syntax as dataset formula metrics); read-only, standalone datasets only, never required, no default, no chaining. Referenced columns cannot be deleted, disabled, or recast while the formula exists (409). See Elastic Data Fields.
  • Elastic Data: Numeric columns of a referenced dataset (integer, decimal, quantity_unit, formula) are now exposed as row-level metrics of the host dataset, keyed <relation>__<column>__value — a relation to a supplier’s emission-factor column can now be used in per-row formulas. The value is the referenced record’s number once per host row (a lookup); SUM counts it per reference. Native entity FK columns lend NUMBER columns the same way.
  • Pivot Table: Added telecommuting_days (total + average) and vehicle_occupancy (total + average) measures to the employee_commuting pivot — telecommuting from weekly_telecommuting days/week, vehicle occupancy from carpool boolean (1 seat solo, 3 when true). Averages roll up correctly at every hierarchy level using hidden SUM/distinct-employee columns. See Pivot Table Config.
  • Pivot Table: Added fuel_kwh and fuel_kwh_financial measures to the fuel-consumption pivot, exposing kWh conversions for stationary/mobile/electricity energy alongside the existing original-unit measures. Added fuel_name_with_unit hierarchy and source_type dimension to the config endpoint.
  • Emission Factors: POST /v1/emission-factors/{category}/config-get-all now accepts custom_values in its body — a {field_key: [values]} map that narrows the listing by the Elastic Data custom columns users add themselves (AND across keys, OR within a key). Only the purchases filter schema declares it; other categories accept the field and ignore it silently rather than erroring. The companion GET /v1/emission-factors/{category}/custom-column-unique-values lists a column’s distinct values for a year — that one does return 400 outside purchases.
  • Invoices: Invoices GET /v1/invoices listing now filters the CO2e impact-sum columns by the custom columns the cells render — only impacts matching the sidecar filters selected for the row are summed.
  • Report Data: Added AI-consolidated group answers for CSRD datapoints — POST /v1/report-data/{datapoint_id}/ai-group-answer returns a synthesized answer from multiple facility/subsidiary responses.
  • Fixed CSRD datapoint listing ordering to respect the sort_order column instead of returning datapoints in insertion order.
  • Fixed invoices GET /v1/invoices/{id} coercing CUPS start_date/end_date to date before response validation — previously failed validation when the database returned datetime.
  • Fixed confidence filter on invoices and waste listings guarding against non-object readings — previously crashed with AttributeError when the reading was a scalar.

September 8, 2026

API · Docs
  • ⚠️ CSRD: Swapped ESRS catalogue to the simplified-2026 standard — sot_csrd_output_data now contains 652 datapoints (down from 1546), and sustainability matters condensed from 90 legacy topics (version esrs_2023) to 37 active topics (version simplified_2026) per EFRAG Appendix A. Existing double-materiality campaigns retain the legacy 90-topic list; new campaigns seed only the 37 active matters. The 27 CSRD auto-dashboard widgets have been remapped to 2026 datapoint codes; 14 widgets with no equivalent were removed. See CSRD.
  • ⚠️ Auth: Removed forced password-expiry policy — login and confirm_mfa_login no longer return password_change_required, and the DAYS_PASSWORD_EXPIRES threshold is gone. Users are never forced to change their password on login. Self-service password change and forgot-password reset are unaffected. See Auth.
  • ⚠️ Rate Limits: Account-wide rate limits (600/min burst, 900/min sustained per API key) are now enforced on dev and staging environments. Production remains in dry-run mode. Requests exceeding the limit return 429 with UPSTREAM_RATE_LIMITED in the response body. The logistics family is exempt from the public baseline. See Rate Limits.
  • Elastic Data: Formula columns may now evaluate to text — the same FormulaNode tree used for numeric formulas can return a text result via if conditionals comparing text columns.
  • Elastic Data: Sidecar formula columns may now reference the category’s own columns in their expressions — a purchases sidecar can reference quantity or unit_price from the purchases table itself, not just borrowed dimensions.
  • Elastic Data: Sidecar formulas may now read the texts behind the row’s foreign keys — if a purchase references a supplier, a formula can access the supplier’s name or country as a text dimension in conditions.
  • Elastic Data: Added defaults in the visualization’s sidecar schema, formula errors as NULL with logging (instead of failing the query), and condition values (for if nodes) exposed on the schema endpoint.
  • Elastic Data: Sidecar per-unit measures are now available in row formulas — widgets can use normalized rates (e.g., emission_factor_per_kg) in custom calculations.
  • Fixed the emission-factor candidate-matches query (GET /v1/emission-factors/{category}/candidate-matches) handling dcycle_refrigerant_gases rows lacking source_id — previously crashed with AttributeError.
  • Fixed GWP conversion-factor deduplication across multi-row base periods — previously summed the same factor multiple times when a substance had multiple base-period rows.
  • Fixed imports anonymous supplier uploads missing processing-job tracking and file association — processing job and file records are now created for the upload.
  • Fixed LCA resource download excluding ACV-derived resource factors — the export now includes all factors.
  • Docs: Published comprehensive rate-limits.mdx reference documenting the account-wide 600/min burst budget, token-bucket refill mechanics, logistics exclusion, and UPSTREAM_RATE_LIMITED response format.

August 2026

August 5, 2026

API · CLI
  • CLI: Added dcy organization-supplier (list/create/edit/delete) for the org-level supplier directory, and dcy ppwr packaging-unit supplier (list/add/remove) to assign PPWR packaging units to a supplier — one supplier per unit, enforced by the API with a 409 on conflict. See Organization Suppliers CLI.
  • API: Hotel stays custom columns now flow through CSV imports too — map an existing sidecar column to a file header, or declare a new one, completing the read/filter/import surface shipped August 4.
  • API: Facilities inbound-email config gains automatic data-reminder emails — PATCH /v1/facilities/{facility_id}/inbound-email/reminder sets a reminder recipient (a free-form email, an existing org member, or both) and an optional threshold in days; a daily sweep emails facilities with no completed inbound data within that window. Setting a recipient email automatically allowlists it, so a plain reply lands back in the same inbound-email pipeline.
  • API: Added cascading delete for Datadis organization identifiers (NIF/CIF) — GET /v1/datadis/organization-identifiers/{nif}/delete-preview shows what would be removed, and DELETE /v1/datadis/organization-identifiers/{nif} cascades to that identifier’s own-org supply contracts and invoices.
  • Fixed vehicle consumption delete endpoints (single, bulk, and bulk-by-filters) authorizing on the read perimeter instead of the write perimeter — a disabled organization’s consumptions were deletable even though its vehicle could not be updated. Cross-organization deletes authenticated with an API key now correctly return 403 (CROSS_ORG_WRITE_REQUIRES_USER) instead of succeeding.
  • Fixed the electricity Scope 1 & 2 report (POST /v2/download/scope-report) always generating one failed processing job alongside the successful one — the endpoint dispatched an unsupported electricity_lb event that the handler has never supported.
  • Fixed PATCH /v1/emission-factors/hotel_stays/config-update blocking the whole batch when the chosen emission-factor source has no coverage for one stay’s country — uncovered stays now keep their existing config and are reported in the processing job’s failed_records instead of failing the request.
  • Fixed the multidb custom emission-factor pipeline further: the primary category was hardcoded to stationary regardless of invoice type, silently dropping the client-choice config for electricity/recharge/water/process invoices using a custom emission-factor database.
  • Fixed purchases bulk-delete-by-filters returning a persistent 409 filter-hash mismatch whenever the preceding list request included a sort parameter — sort no longer affects the filter hash, matching that it only controls display order.
  • Fixed imports auto-mapping for Spanish waste ledgers: 4-digit MITECO treatment subcodes (e.g. R1201) now resolve to their parent operation code instead of falling through to manual mapping.
  • Fixed invoices and wastes custom-column creation accepting reserved grid keys (edit, select, toc) — declaring a custom column with one of these names silently broke the shared facility-detail grid’s action column; it’s now rejected with 422.
  • Fixed POST /pivot_table/calculate sorting Scope 3 GHG categories lexicographically by label instead of numerically (Category 11 sorted before Category 2) — fixed backend-side so on-screen tables, Excel, and PDF exports are all consistent.
  • Fixed invoice create/update silently skipping the emission-factor config when a linked custom EF record had no available source for the invoice’s primary category — now fails fast with a 422 (MULTIDB_CUSTOM_RECORD_INCOMPATIBLE_CATEGORY) instead of falling back to standard emissions unannounced.
  • Fixed the vehicles sheet in the Scope 1 & 2 report (POST /v2/download/scope-report) shipping empty — a type mismatch on currency_unit_id raised on every row, and the export silently dropped each one rather than failing the job.
  • Fixed LCA transport (A2) impacts calculating as 0 for portfolios whose transport was created via the API or older UI flows as a legacy “copy-piece” — those are now bridged into the current chain/leg model instead of being invisible to the calculation.
  • Raised the facility_consumption import quantity limit from 50M to 10B.
  • Fixed hotel stay emission-factor configs only being created for a stay’s check-in year — a stay spanning a year boundary (e.g. Dec 28 – Jan 3) now gets configs for every year it touches.

August 4, 2026

API · CLI
  • CLI: Added dcy ppwr packaging-unit (list/show/create/edit/delete) and dcy ppwr packaging-unit sku (list/add/edit/delete) subcommands for managing PPWR packaging units and the SKUs they cover directly from the CLI.
  • CLI: dcy file list now shows a linked_packaging_units column — which PPWR packaging unit(s), if any, were created/applied from that file.
  • API: LCA calculation errors now also surface four transport (A2) data problems — missing_transport_destination, missing_transport_origin, missing_transport_data, and missing_transport_emission_factor — mirroring the silent-omission checks already in place for materials (A1).
  • API: Vehicle consumptions gain custom columns support — custom_values is now returned on list, org-wide list, and detail responses, matching the existing purchases/employees/facilities behavior. POST /v2/vehicle_consumptions/migrate now returns 409 instead of silently dropping an org’s custom values on a cross-organization move.
  • API: Hotel stays gain custom columns support — custom_values on responses and custom_column_schemas on the list page, plus column-header filtering (filter_by/filter_match/filter_or_fields) on GET /hotel-stays, matching the filtering already available on purchases.
  • API: Transport routes gain custom columns support — custom_values on list/detail responses, sortable via sort_by, and mappable/declarable through CSV imports.
  • API: employee_commuting imports now support custom columns — map an existing employees_historic sidecar field to a file header, or declare a new one, through the imports API.
  • API: Purchases list now returns custom_column_schemas under group view, matching vehicles/vehicle-consumptions/hotel-stays/business-travels — a subsidiary’s custom column now renders under its own label instead of the holding’s.
  • Fixed cross-tenant authorization gaps on transport (climate) plan endpoints — reading, updating, or deleting another organization’s plan or measure is no longer possible for an authenticated caller outside that organization’s family.
  • Fixed transport routes sort_by accepting any model attribute name (including relationships) and returning a 500 instead of only real, orderable columns.
  • Fixed PPWR packaging unit detail/update/delete endpoints (GET/PATCH/DELETE /v2/ppwr/packaging-units/{id}) to enforce organization scoping, matching the existing SKU and substance-declaration sub-resource behavior.
  • Fixed a malformed x-organization-id header (non-UUID) on API-key authenticated requests returning a 500 instead of a 401.
  • Fixed imports of the facility_consumption electricity/heat/water/recharge sub-templates silently dropping mapped or declared custom columns instead of persisting them.
  • Fixed a 500 error when submitting employee_survey_emails imports through the imports API.
  • Fixed the Scope 1 & 2 electricity/energy report (POST /v2/download/scope-report) undercounting emissions for invoices whose billing period straddles the reporting-window boundary.
  • Fixed the multidb custom emission-factor pipeline: 3-gas (CO2/CH4/N2O) custom factors weren’t converting to CO2e (missing GWP conversion config), and only the primary category was processed for electricity/heat invoices, leaving secondary categories (location-based, generation, T&D) uncalculated. start_date on custom emission-factor versions (POST /v1/emission-factors/custom-databases/{database_id}/versions) is now optional, defaulting to 1970-01-01 like standard conversion factors.
  • Fixed custom-column filtering and sorting under group view (consolidate_group) reading every row’s value from the header organization’s sidecar instead of the row’s own — a subsidiary’s custom-column value could silently drop out of a filtered result or sort as if empty. Affects purchases, vehicles, vehicle consumptions, business travels, facilities, wastes, hotel stays, invoices, employees, suppliers, and own workforce trainings.
  • Fixed dropped custom-column values during CSV imports going unreported — now surfaced via the processing job’s job_metadata.failed_records, the same channel the Activity → Processing Jobs detail view already renders.

July 2026

July 27, 2026

API
  • ⚠️ API: DELETE /v2/elastic-data/datasets/{dataset_id} no longer returns 204 No Content — it now returns 200 with a processing-job reference body, tracking the deletion asynchronously. See Elastic Data.
  • ⚠️ API: Purchases create endpoints drop the multidb query param — all organizations now route through the v2 conversion-factors pipeline unconditionally.
  • API: Elastic Data text→date/datetime column recasting now accepts day-first DD/MM/YYYY (in addition to ISO YYYY-MM-DD), matching what the UI displays, and surfaces the expected format on a bad value. See Elastic Data.
  • Fixed the project task versions endpoint — version_changes (what changed between versions) was never populated, and the most recent version was silently dropped from the list.
  • Fixed dataset visualization widgets rendering empty in project reports and their DOCX/PDF exports — they now resolve through the same query path as POST /datasets/{key}/query.

July 24, 2026

API · MCP
  • MCP: Expanded Elastic Dataset tools — query_dataset gains sort and filters params; added get_dataset_record, list_dataset_groups (dataset folders), and list_project_datasets tools.
  • API: Facilities update and delete now authorize across the whole accepted business family instead of the header org only — subsidiary facility rows are editable/deletable from the holding, mirroring how sold_products already works. Facility creation stays header-org only.
  • API: Logistic hub update/delete get the same family-wide authorization as facilities — subsidiary hub rows are editable/deletable under group view.
  • API: Transport (shipments/deliveries) bulk delete and own workforce bulk delete now follow the group view — family-scoped when consolidate_group is set, instead of only the header org.
  • API: Added Greenview hotel-stay emission factors for 227 additional countries (plus a missing Serbia match), broadening hotel stay CO2e coverage.
  • Fixed LCA dashboard calculation errors listing the same underlying error multiple times per piece; block-scoped error codes are still preserved.
  • Fixed OR-matching in Elastic Dataset record filters returning incorrect results.

July 23, 2026

API
  • API: Vehicle consumption bulk delete (by IDs and by filters) now operates across the whole organization family when consolidate_group is set, instead of only the header org. Default behavior unchanged.

July 22, 2026

API
  • ⚠️ API: GET /v1/logistics/clients no longer accepts project_id (the client list is now organization-wide, not project-scoped) and is now keyset-paginated via new search, last_client_name, and limit params — reads a dedicated lookup table instead of scanning the full logistics-requests history, fixing timeouts for large clients.
  • API: Added Scope 3 Category 13 (downstream leased assets) as a new GHG emission category (downstream_leased) alongside the existing upstream-leased-assets category — accepted as a category value on POST /historical-emissions and in validated_ti_all/get_total_impacts/get_greenhouse_gas_emissions filters.
  • API: Facilities list gains the same opt-in group view as business travels, vehicles, logistic hubs, transport, own workforce, and employees — a consolidate_group param for holding-wide data.
  • API: LCA block responses now include updated_by_name and updated_by_logo_url for the user who last updated a linked custom transport route.
  • API: Purchases list, totals, and count gain the same opt-in consolidate_group group view as the other modules — spans the org + descendants, each row enriched with organization_id/name/logo.
  • API: Own workforce trainings paginated list gains the same opt-in consolidate_group group view as the accidents & absences list.
  • API: Employees GET /v2/employees gains a survey_template_id filter and now returns it in the response.
  • API: The emission-factors detail export now accepts start_date/end_date in addition to year.
  • Fixed the vehicle bulk-upload fuel-type dropdown excluding battery_electric_vehicle and plug_in_hybrid_electric_vehicle, plus a related vehicle-type/fuel compatibility validation bug.
  • Fixed DELETE /v1/projects/{project_id} still failing when the project had import sessions attached — sessions are now cleaned up first instead of only returning a 409.
  • Fixed employee commuting CO2e (Scope 3 Cat. 7) to align with total_impacts values, removing synthetic walking-employee rows that no longer matched the underlying calculation.
  • Fixed purchases bulk imports for v2 organizations — D&B supplier enrichment previously skipped them, leaving purchases unmatched to emission-factor configs and product_name empty.
  • Fixed CSV/XLSX header-row detection for files with sparse columns (filled only in the header row, never in data rows), which could misdetect the header on certain imports.
  • Fixed purchases imports to resolve custom emission-factor groups and units defined on a parent organization — previously only descendant orgs were searched, blocking child-org imports.
  • Fixed license plate matching during multi-org CSV imports to ignore hyphen/space formatting differences (normalizes to alphanumeric-only before comparing).
  • Fixed heat invoice impact calculation falling through with no result when the v2 pipeline had no conversion steps — now falls back to the legacy emission-factor columns.
  • Fixed the LCA scenario materials import template forcing material, supplier, and country into one combined dropdown value — the material_id column is now material_name, and each of the three now maps and validates independently.
  • API: Added a product_name[] filter to the purchases list.
  • API: Purchase suppliers list (GET /v1/purchases/suppliers) gains the same opt-in consolidate_group group view as the other modules.
  • API: Vehicle bulk delete (by IDs and by filters) now operates across the whole organization family when consolidate_group is set, instead of silently dropping subsidiary rows. Default behavior unchanged.
  • Fixed the business travel download export showing an emission factor from a frozen reference table instead of the one the calculation engine actually applied, so the exported factor and CO2e didn’t reconcile — the export now sources the factor and calculation chain from each trip’s own conversion_steps, and adds a conversion_steps_detail column showing the full multi-gas chain.

July 21, 2026

API
  • API: PPWR packaging units gain an explicit ppwr_category field (Regulation (EU) 2025/40 Art. 3.1: sales/grouped/transport/e_commerce) — a user-set classification, never inferred, that now drives the gap and compliance engines instead of the legacy packaging_type. Also importable via a unit_ppwr_category CSV catalog column.
  • API: PPWR gap analysis adds category-conditioned obligations: Art. 24(1) void-space ratio cap for grouped/transport packaging (new void_space_pct field), Art. 25 + Annex V single-use plastic ban for grouped packaging (new is_single_use field), and Art. 29 reuse-target advisories.
  • API: Added POST /v2/ppwr/packaging-units/bulk-delete for organization-scoped batch deletion of packaging units, and a search query param on the packaging units list endpoint for case-insensitive filtering by unit name or SKU reference.
  • API: Own workforce accidents & absences list and employee mobility surveys list (GET /v2/employees) gain the same opt-in group view as business travels, vehicles, logistic hubs, and transport — a consolidate_group param plus organization_id[] filter for holding-wide data.
  • API: Added an opt-in scope_to_project_organizations query param to the extended projects list — when true, projects owned by a parent/ancestor organization are only returned if the caller’s org is a project_organizations member of that project. Defaults to false, unchanged behavior.
  • API: The logistics report export now includes vehicle_license_plate and trailer_license_plate columns.
  • API: Added a bulk emission-factors export endpoint — queues an async export of emission-factor-detail records by organization(s), year, and category, with an export_by_company option to produce a per-company ZIP instead of a combined workbook.
  • API: The business travel download export now includes the passenger_number column (previously computed but omitted from the file).
  • API: Fixed integer fields above 2^53 being silently rounded during dataset ingestion — bare integer strings are now parsed exactly instead of routed through float().
  • API: Imports — custom-dataset column auto-mapping now recognizes a file header that contains every token of the field name plus an extra descriptive suffix (e.g. “Salario bruto oficina central” for field “Salario bruto”), reducing manual remapping. See suggest-mapping.
  • Fixed incorrect PPWR Art. 7 recycled-content targets — the gap engine and the Declaration of Conformity resolver had disagreed with each other and with the regulation; both now read from one constants module.
  • Fixed the PPWR supplier-survey template generator silently dropping questions for previously unmapped gap codes (e.g. HEAVY_METALS_NON_COMPLIANT); its gaps_addressed count now matches the questions actually generated.
  • Fixed employee emissions calculated via the legacy path sometimes showing co2e=0 in GET /v2/employees responses even though the underlying emissions were non-zero.

July 20, 2026

API
  • API: Added the Task Completion Assistant — POST /v2/task-completion/projects/{project_id}/evaluate streams (SSE) an AI-assisted evaluation of which project tasks are ready to be marked complete, combining deterministic data-presence signals with an LLM re-ask for ambiguous cases; POST /v2/task-completion/projects/{project_id}/complete bulk-completes the selected tasks behind a server-side data-quality gate. Both support scoping to a subsidiary via organization_id and localized messages via Accept-Language.
  • API: Logistic hubs list gains the same opt-in group view as business travels and vehicles — a consolidate_group param (org + accepted descendants, each hub carrying its owning organization) and an organization_id[] filter to scope it further.
  • API: Added an optional request_date field (YYYY-MM-DD) to POST /v1/logistics/requests and its bulk variant — distinct from shipment_date, for customers who cross-reference logistics data against their own order-request date.
  • API: Employee survey custom fields support conditional branching — new conditions (show child fields when a given option is selected) and depends_on (link a field to a built-in step’s condition group) on CustomFieldSch.
  • API: Added bulk invoice splitting — POST /v1/invoices/bulk-split-preview, bulk-split (by IDs), and bulk-split-by-filters. See invoices.
  • API: Transport (shipments/deliveries) list gains the same opt-in group view as business travels, vehicles, and logistic hubs — a consolidate_group param with group-aware totals and an organization_id[] filter.
  • Fixed vehicle consumption bulk uploads and CSV imports mishandling plates shared across organizations (fleet movements): auto-routing to the correct subsidiary now falls back correctly instead of rejecting the whole batch, and a plate genuinely ambiguous across multiple active orgs now requires an explicit organization instead of guessing one.
  • Fixed a heat invoice impact miscalculation caused by an untyped UUID from a DB query; added missing standard gross-GJ conversion factors.
  • Fixed DELETE /v1/projects/{project_id} returning a 500 instead of a 409 when the project still has import sessions attached.
  • Fixed ROAD/RAIL logistics requests between separate landmasses (e.g. mainland Spain ↔ Canary Islands, Ibiza ↔ Mallorca) silently returning a near-zero, physically impossible distance instead of routing as maritime.

July 15, 2026

API · CLI
  • API: Custom columns are now supported on eight more categories — employees, wastes, vehicles, business travels, invoices, invoice suppliers, facilities, and waste water treatments — matching the existing purchases behavior, with custom_values now returned on list/detail responses. Employees, wastes, business travels, and invoices also include custom columns in their downloaded exports.
  • API: Business travels and sold products list endpoints gain an opt-in consolidate_group query param for a group-wide view across the header organization and its descendants — responses add organization_name/organization_logo_url. Defaults to false, so existing callers are unaffected.
  • API: Added GET /v1/company-lookup/match, /enrich, and /logo — a two-step Dun & Bradstreet company lookup (cheap name+country match, then a billable data enrich step) plus a Brandfetch logo suggestion, for the freemium onboarding flow. Restricted to accounts still in onboarding.
  • API: GET /purchases/{id}/steps now returns an explicit sources array per calculation period (source id, name, version), making calculation provenance explicit instead of implicit.
  • API: Added non_currency_purchases to the purchases unit-type enum, plus new ecoinvent unit options (kWh, MJ, ha) for custom emission factors.
  • API: Wastes gain a provider_name field (OCR-extracted service-provider name) on create, update, list, and detail.
  • API: Projects gain a status field (active/archived); the extended projects list supports filtering by it.
  • API: LCA imports — direct material-to-process links now create connectable blocks matching the manual canvas; import summary counts (total_inputs/total_output/total_blocks) are fixed from always returning zero; the dashboard now surfaces validation_errors for pieces that can’t compute; incompatible units raise a clear INCOMPATIBLE_UNIT error instead of silently corrupting piece data.
  • API: Imports — category-column value mappings are now remembered per organization/template across sessions; mapping/suggest responses add history_restored_count/stale_history_mapping_count.
  • API: Imports — vehicle bulk-upload fuel-type options now include electric, hybrid, and electric_renewable; large workbooks no longer error on mapping/suggest (embedding requests are chunked under the model’s item limit).
  • API: Facilities creation now accepts address_place_id/address_latitude/address_longitude from a Google Places result, skipping strict address validation that previously rejected valid rural/polígono addresses.
  • CLI: Added dcy lca delete-material and dcy lca bulk-delete-materials (alias rm-materials) subcommands. See LCA CLI.
  • Fixed the v2 supplier-code PATCH endpoint storing the wrong mapping ID for non-SIC (CNAE/NACE) codes, causing the assigned code to disappear on the next GET; also fixed a 404 when assigning CNAE/NACE codes, since lookups only checked the SIC table.
  • Fixed ChangePassword.old_password incorrectly enforcing an 8-character minimum, which raised an unhandled error for accounts with a legacy shorter password.
  • Fixed CustomEmissionFactorDataEnergy.renewable_percentage — posting a custom energy emission factor without it no longer errors.

July 14, 2026

API · CLI
  • API: Added GET /v2/download/scope-report/{combustion,electricity,vehicles,recharge} per-category scope-report endpoints; Excel exports gain total_spend and currency columns.
  • API: Facilities list/detail responses now expose an inbound_email status field for the facility’s inbound-email onboarding state.
  • API: Imports — session creation accepts new routing-context parameters (target_organization_id, measurement_campaign_id, lca_id, target_block_id) and a new anonymous suppliers import category. See imports.
  • API: Elastic Data — datasets can now be organized into folders (groups): new group CRUD endpoints and a group_id field on datasets. See datasets.
  • API: Elastic Data — schema inference (POST /v2/elastic-data/inferences) accepts an optional dataset_id to evolve an existing dataset’s schema instead of replacing it.
  • CLI: Fixed dcy purchase supplier commands — set-code, list, and export had incorrect download-path routing, broken pagination across all matches, and country-matching bugs; CSV parsing and export are also hardened against ragged rows and formula injection.
  • Fixed GET /v2/vehicle_consumptions and the facility wastes list returning a server error on an invalid sort value — both now fall back to the default order; the wastes list also gains previously-ignored sortable columns (description, total_km_to_waste_center, identification_name, updated_at).
  • Fixed PATCH /v1/purchases/{id} (and the vehicle/invoice equivalents) crashing on a valid custom_emission_factor_id due to a UUID/string type mismatch.
  • Fixed a spurious validation error on emission-factor config updates when record_ids contained duplicates.

July 13, 2026

API
  • API: Added a native Category 11 electricity audit export for sold products — a per product × country × period CSV with the emission-factor formula, source, and database used.
  • API: PPWR gap-analysis endpoints (/v2/ppwr/) now localize gap descriptions and required actions via the Accept-Language header (Spanish supported, English default).
  • Fixed a geocoding bug where addresses ending in an ISO alpha-2 country code (e.g. “Barcelona, ES”) could resolve to the wrong country, producing incorrect distances for business travel, logistics, and facility data.
  • Fixed exports for custom emission factors showing “Unknown” instead of the emission-factor group name in the Provider column.
  • Fixed the employees “unique values” endpoint not surfacing telecommuting as a work_model value alongside hybrid.

July 10, 2026

API · MCP
  • MCP: Added create_vehicle, update_vehicle, create_employee, and update_employee write tools for managing vehicles and employees programmatically — supports type, ownership, license plate, country, situation, status, and partial updates.
  • MCP: Added create_invoice and update_invoice write tools for managing invoices programmatically, with type and status validation.
  • MCP: Added create_logistics_request write tool for creating logistics requests programmatically, with category, TOC, load unit, and subcontractor type validation.
  • MCP: Added get_workforce_contracts read tool — exposes employee contract history filtered by own_workforce_id, returning start/end dates for contract period tracking. See workforce.
  • MCP: Added get_workforce_remunerations tool to retrieve salary and compensation records for a given workforce contract, including amount, currency, and period dates. Completes the workforce tool set.
  • MCP: Added list_double_materiality tool — retrieve CSRD Double Materiality assessment topics with optional year, lang, and material filters (docs).
  • MCP: Added list_csrd_iros tool — retrieve CSRD IROs (Impacts, Risks, and Opportunities) for a given year (docs).
  • MCP: Added get_csrd_output_data tool — retrieve ESRS disclosure requirements grouped by code, with optional date range filters (docs).
  • MCP: Added list_measurement_campaigns tool — retrieve measurement campaigns (reporting periods) for an organization (docs).
  • MCP: Added list_csrd_survey_results tool — retrieve CSRD stakeholder survey results with impact and financial materiality scores, filterable by year and lang (docs).
  • MCP: Added 6 PPWR packaging compliance tools — list_packaging_units, get_packaging_unit, get_packaging_compliance, get_project_compliance, get_packaging_gaps, and get_ppwr_dashboard. Evaluate EU packaging regulation compliance at unit and project level, identify gaps, and view dashboard metrics. All use /v2/ppwr/ routes. See PPWR.
  • API: Added optional surface_area (numeric) and surface_area_unit (string) fields to facility create, update, and read schemas (/v2/facilities/). Also available on waste water facility creation. Both fields are version-tracked.
  • MCP: Added list_waste_water_treatments and get_waste_water_treatment tools — query Scope 1 waste water treatment records by facility with date range, status, and CO2e status filters. See waste water.
  • MCP: Added list_sold_products, get_sold_product_periods, and get_sold_product_period_emissions tools — access Scope 3 Category 11 (use of sold products) data including product catalog, reporting periods, and per-period emission breakdowns. See sold products.
  • Fixed POST /v1/employees/csv rejecting multi-column CSVs (e.g. Organization + Email) with a FILE_PARSE_ERROR — the CSV branch now picks the email column the same way the Excel branch does.

July 9, 2026

MCP
  • MCP: Added five new write tools for Scope 3 data entry — create_purchase (purchased goods & services), create_business_travel (employee travel with origin/destination or manual distance), create_hotel_stay (accommodation with nights or check_out_date), create_transport (multi-leg shipments with per-section transport mode), and create_waste (waste generated in operations — Scope 3 Category 5). All include client-side enum validation and async CO2e calculation.
  • MCP: Added update_purchase write tool for partial updates to existing purchases — supports product_name, quantity, unit_id, purchase_date, expense_type, description, recycled, and custom_emission_factor_id fields with client-side expense_type validation.
  • MCP: Added update_business_travel write tool for partial updates to business travel records — supports transport_type, dates, origin/destination, distance_km, vehicle details, and trip metadata with client-side enum validation.
  • MCP: Added update_hotel_stay write tool for partial updates to hotel stay records — supports dates, country, hotel name/address, and room count with automatic CO2e recalculation.
  • MCP: Added update_transport write tool for replacing transport records — accepts a full sections array with per-leg transport mode, validates transport_type, travel_method, and frequency enums.
  • MCP: Added update_waste write tool for partial updates to waste records (Scope 3 Category 5) — supports identification_name, base_quantity, dates, destination, total_km_to_waste_center, and emission factor fields with automatic CO2e recalculation.
  • MCP: All Scope 3 write tools are now complete — 5 create tools and 5 update tools covering purchases, business travels, hotel stays, transports, and wastes.
  • MCP: Added create_decarb_target write tool for creating reduction targets within decarbonization plans — supports scope (1-3), target_type (absolute/intensity), base_year, target_year, base_emissions, and reduction_percentage with client-side validation.
  • MCP: Added create_decarb_action write tool for creating concrete decarbonization initiatives linked to reduction targets — supports title, target_id, estimated_co2e_reduction, estimated_cost, payback_period_months, priority (quick_win/medium_term/long_term), and initiative_type with client-side validation.
  • MCP: Added create_project write tool for creating ESG projects — supports name, responsible_user_id, project_type (17 framework types including carbon_footprint, iso_14064, lca, esrs), methodology, and date fields with client-side validation.
  • MCP: Added create_project_task write tool for creating tasks within projects — supports title, assigned_to, due_date, stage (pending/in_progress/completed), category, tags, and subtask nesting via parent_task_id with client-side stage validation.
  • MCP: Added submit_kpi_response write tool for submitting values to Custom KPI campaigns — supports numeric, text, date, and select-type KPIs with optional facility_id scoping. Each entry requires a kpi_id and the matching value field (value_numeric, value_text, value_date, or option_id).
  • MCP: Added data entry workflow guidance to the server’s _INSTRUCTIONS prompt — LLMs now receive step-by-step chains for each write tool and a reminder to confirm with users before creating records.
  • MCP: Added list_workforce read tool — the first tool covering the ESG social pillar (own workforce / employees). Exposes demographics, employment categories, contract dates, locations, and nationality. See workforce.
  • MCP: Added list_workforce_absenteeisms read tool — exposes workforce absenteeism and workplace accident records. Each record is either an absenteeism (hours lost) or an accident (date + context), covering ESRS health & safety social disclosures. See workforce.
  • MCP: Added list_workforce_trainings read tool — exposes workforce training records (type, hours, dates, description) for ESG social disclosures. See workforce.

July 8, 2026

API · MCP
  • API: Added nights parameter to hotel stay create and update endpoints as an alternative to check_out_date — when nights is provided with check_in_date, the checkout date is derived automatically. Either check_out_date or nights must be supplied on creation.
  • API: Added custom_values response field to purchase list, detail, create, and update endpoints — returns elastic-data custom column values as a key-value dictionary. Purchase list sorting now also supports custom column keys.
  • MCP: Added create_facility and update_facility as the first write tools for the MCP server — AI assistants can now create facilities with all standard fields and update existing ones with partial payloads and client-side status validation.
  • API: Added GET /v2/elastic-data/datasets/{dataset_id}/records/{record_id} endpoint to retrieve individual dataset records by ID.

July 7, 2026

API · CLI · Documentation
  • API: Added consolidated reporting endpoints — POST /v1/total-impacts/purchases/report returns product-level purchase emissions ranked by CO2e, and POST /v1/total-impacts/hotel-stays returns hotel-stay emissions (Scope 3 Category 6) with monthly buckets and weighted totals.
  • API: Added header row auto-detection for tabular imports — the header_row parameter and header_row_candidates response field let API consumers handle files with title rows above the actual column headers.
  • ⚠️ CLI: Resolved flag shorthand collisions — removed -y from --year/--registration-year (now means --yes) and -v from --verbose on vehicle diagnose/upload (now means HTTP debug). Use the long forms --year and --verbose instead.
  • CLI: Extended shell tab-completion to purchase, shipment, widget, import, and file-reading commands — widget supports two-level completion (project ID → widget ID). Also auto-registered --project-id and --file-id flag completions across all commands that accept them.
  • CLI: Routed all human-facing messages (cancellation notices, empty-result warnings, validation errors, verbose output) to stderr — --json output on stdout is now reliably parseable via pipes.
  • Documentation: Restructured CLI docs — split large pages into standalone command pages for facilities, vehicles, employees, invoices, purchases, transport, and 10+ more commands.

July 6, 2026

API · CLI · MCP
  • Added land use change (land_use_change) as a new GHG emission category across API, MCP, and historical emissions (Scope 1 — LULUCF). The MCP get_greenhouse_gas_emissions tool now accepts category='land_use_change'.
  • CLI: Added shell tab-completion for positional arguments on facility, vehicle, employee, file, project, hub, transport, and member commands — IDs are suggested with human-readable names or license plates.
  • CLI: Changed invoice list --size default from 100 to 50 to align with all other list commands.
  • MCP: Fixed prompt templates (carbon_footprint_summary, fleet_report, logistics_overview) to default year to the previous calendar year instead of hardcoded 2024.

July 5, 2026

CLI
  • CLI: Added SIGTERM handling for clean shutdown in containers — the CLI now exits gracefully with code 143 when stopped by Docker, Kubernetes, or systemd.
  • Documentation: Expanded CLI docs with exit codes reference, structured JSON error output format, auth status --check for credential validation, automatic retry behavior, --json shorthand, dcy doctor exit codes, and config unset/set validation.

July 4, 2026

MCP · Documentation
  • MCP: Added client-side enum validation to 14 tools — catches invalid status, ownership, situation, transport_type, category, framework, impact_category, and other enum parameters before API calls with clear error messages instead of silent empty results.
  • MCP: Enriched docstrings for 9 tools with GHG scope context and usage hints, improving LLM tool discovery for list_vehicles, list_employees, list_transports, list_files, get_recent_activities, list_logistics_hubs, get_logistics_report, get_custom_kpi, and get_kpi_campaign_values.
  • MCP: Fixed incorrect enum values in vehicles, hotel stays, transports, and business travels tool documentation.
  • MCP: Added Gemini CLI configuration to the MCP quickstart guide.
  • Documentation: API response field accuracy sweep — fixed types, nullability, and missing fields across 20+ endpoint pages. Added cross-links between API, CLI, and MCP documentation.

July 3, 2026

API · MCP · Documentation
  • API: Added async logistics report export endpoint GET /v1/logistics/report/export — queues a background job to generate a detailed row-per-leg export (CSV, Excel, or JSON with optional per-client ZIP) and emails a download link when ready.
  • MCP: Added initial client-side enum validation and enriched tool docstrings for better LLM self-correction and tool selection.
  • MCP: Expanded _INSTRUCTIONS with workflow guidance and best practices for multi-step analysis.
  • Documentation: Removed 56 internal/UI-only endpoint pages from public API docs — the developer center now shows only endpoints useful for external API consumers.
  • Documentation: Fixed API base URLs from /api/v1/ to /v1/ and improved the API introduction with versioning and rate limit information.
  • Documentation: Enriched CLI documentation across all command pages with output examples, flag tables, JSON examples, and missing subcommands.
  • Documentation: Elevated 8+ MCP tool pages with response examples, parameter fixes, and scope guidance.

July 2, 2026

Documentation
  • Documentation: Added Common Errors sections to all API endpoint pages with error codes, descriptions, and resolution guidance.
  • Documentation: Added LCA endpoint pages — create-from-json, create-from-excel, validate, impacts, export-xlsx, and dashboard.
  • Documentation: Added imports endpoint pages for column-values and error-groups.

July 1, 2026

Documentation

June 2026

June 30, 2026

API · MCP
  • MCP: Expanded toolset from 21 to 45 tools — added 24 new read-only tools across facilities detail, wastes, hotel stays, transports, logistics (clients/hubs/report/summary), files, historical emissions, total impacts, emission factors, projects, dashboards, decarbonization plans, KPI datasets/campaigns/assignments, LCA block/piece impacts + validation, elastic data, and compliance.
  • API: Added Elastic Data field/record CRUD endpoints with type-aware filtering and sorting — new POST/PATCH/DELETE for fields and records with operator grammar for typed queries.
  • API: Added LCA ecoinvent activity linking — three new endpoints for searching ecoinvent 3.11 activities (GET /v1/lca/ecoinvent/activities), resolving locations (GET /v1/lca/ecoinvent/activity-locations), and linking materials to impact sources (POST /v1/lca/materials/{id}/link).

June 29, 2026

API · MCP
  • API: Exposed LCA portfolio and dashboard via public API — GET /lca/portfolio and GET /lca/portfolio/{lca_id}/dashboard now accept API key authentication alongside JWT.
  • MCP: Added LCA tools — list_lcas for querying portfolios and get_lca_dashboard for EN 15804 environmental impact results.

June 15, 2026

API
  • API: Added employees bulk delete endpoints — delete by IDs (POST /v1/employees/bulk-delete) and by filters (POST /v1/employees/bulk-delete-by-filters).

June 8, 2026

API
  • API: Added custom KPI Excel download endpoint (GET /v1/custom-kpi/datasets/{id}/download) for programmatic data export.
  • API: Added GET /v1/invoices/totals endpoint for aggregated invoice metrics.
  • API: Added filter_by parameter to GET /v1/purchases for server-side per-column filtering.

June 1, 2026

MCP
  • MCP: Fixed Clerk SSO token support in the MCP OAuth authentication flow.

May 2026

May 9, 2026

API · Documentation
  • API: Added logistics fuel recharges endpoints — POST /v1/logistics/recharges and POST /v1/logistics/recharges/bulk for recording fleet fuel consumption.
  • Documentation: Added Security & Compliance page covering API key lifecycle, data protection, access control, and incident response.

May 5, 2026

MCP

May 4, 2026

Documentation

April 2026

April 23, 2026

API
  • API: Added POST /v1/invoices/recalculate for async invoice recalculation with email notification.
  • API: Added new filter parameters and filter-options endpoint to invoices.

April 22, 2026

API · MCP
  • API: Exposed total_energy_kwh on vehicle consumption endpoints.
  • API: Added GET /v1/waste-efs endpoint for querying waste emission factors.
  • MCP: Added get_greenhouse_gas_emissions tool to MCP emissions page.

April 17, 2026

API
  • API: Added wastes list, detail, bulk-delete, and unique-values endpoints.

April 16, 2026

CLI · Documentation
  • ⚠️ CLI: Renamed CLI binary from dcycle to dcy — all commands now use the dcy prefix.
  • Documentation: Completed dcy CLI command coverage in developer center docs.
  • Documentation: Added ISO 14001 management systems guide.

April 8, 2026

API · MCP
  • API: Added Imports module — GET /v2/imports/sessions for listing and filtering import sessions, plus provider-backed options endpoint.
  • MCP: Added remote mode setup for VS Code, ChatGPT, and Copilot Studio to MCP quickstart.

April 7, 2026

Documentation
  • Documentation: Added Imports section to the developer center.

April 6, 2026

MCP
  • MCP: Added Copilot Studio, ChatGPT, and Gemini CLI configuration guides to MCP quickstart.

April 1, 2026

MCP
  • MCP: Launched remote MCP server at mcp.dcycle.io — no local installation required.

March 2026

March 31, 2026

API
  • API: Added autonomous file upload flow — POST /v1/files/upload and POST /v1/files/process for programmatic document upload with AI-powered extraction and retry support.

January 2026

January 30, 2026

CLI · MCP · Documentation

January 12, 2026

Documentation
  • Documentation: Added ISO 14064 documentation structure with category-by-category emission guides.

January 11, 2026

API
  • API: Updated API key authentication to include user attribution on API key requests.

December 2025

December 23, 2025

API

December 11, 2025

API

December 3, 2025

Documentation
  • Documentation: Added GHG Protocol Scope 3 guides — Category 1 through Category 9, covering purchased goods, capital goods, fuel & energy, upstream transport, waste, business travel, employee commuting, and downstream transport.

December 2, 2025

API

November 2025

November 20, 2025

API · Documentation