Collate AI & Credits

AskCollate is renamed to Collate AI

Breaking for anything that matches on the product name. AskCollate is renamed to Collate AI everywhere: the UI, the backend, billing labels, chat prompts, and the Slack and Microsoft Teams integrations.

Update saved searches, dashboards, runbooks, alert rules and Slack/Teams workflow triggers that match on the string AskCollate. Billing exports change label too, so any spreadsheet or BI report keyed on the old label needs repointing.

Conversations also become collaborative, and the credit model changes from a flat token count to a weighted, cache-aware calculation.

Credits are now weighted

Behavioural. The same workload consumes a different number of credits in 2.0 than in 1.13.

1.13 charged credits from a flat token count. 2.0 introduces a weighted, cache-aware model:

Change	Effect
Weighted by token type	Input, output, cache-read and cache-write tokens carry different cost ratios, so a cache-heavy conversation costs materially less than the same token count of fresh generation
Weighted by model	Credits are weighted by the model that actually served each call, so a request routed to a cheaper model costs fewer credits
Fixed divisor	The weighted basis is divided by a locked `tokensPerCredit` of 58,000
Cache-only usage is recorded	Lump-sum recording is guarded on the weighted basis so cache-only usage is no longer silently dropped
Per-conversation attribution	Credits are attributed to the conversation that spent them, not just to the tenant pool

If you forecast spend from 1.13 token counts, re-baseline after upgrading. Re-run a representative week and compare, rather than converting the old numbers.

Runs that cannot finish are refused

Behavioural. 2.0 warns before the credit limit is reached and refuses to start work that cannot complete within the remaining pool.

In 1.13 a long automation could start and exhaust the pool part-way through, leaving partial results and no credits. In 2.0 the run is rejected up front with a clear reason. Operators get a live view of the pool at GET /v1/status/credits, and credits can be refunded from the operator API.

Billing is redesigned

The Billing area is reorganised into Overview, Product, User Adoption, AI Credits, and MCP requests.

Collate 2.0 Billing area showing Overview, Product, User Adoption, AI Credits and MCP requests tabs

New endpoint	Purpose
`GET /v1/billing/credits/conversations`	Per-conversation credit consumption, paginated
`GET /v1/billing/adoption/users`	Per-user adoption
`GET /v1/billing/weekly-impact`	Weekly impact digest data
`POST /v1/billing/weekly-impact/send-digest`	Send the digest, with a date-aware subject and header

An MCP requests usage section is added, backed by a daily MCP usage rollup in the telemetry payload, and AI chat response time is now measured server-side rather than in the browser.

Collate 2.0 Billing MCP requests usage section

Admins can read any conversation from Billing in read-only mode, so that credit spend can be explained. This is a new capability. If your policy forbids administrators reading user conversations, restrict the admin role accordingly before upgrading.

Conversations become collaborative

Additive. A conversation is no longer private to its author.

New endpoint	Purpose
`PUT /v1/assistants/chatConversations/{id}/collaborators/{userId}`	Add a collaborator
`DELETE /v1/assistants/chatConversations/{id}/collaborators/{userId}`	Remove a collaborator
`PUT .../collaborators/{userId}/role`	Change a collaborator’s role
`GET /v1/assistants/chatConversations/{id}/collaborators`	List collaborators
`GET /v1/assistants/chatConversations/collaborators/search`	Search users to add (bots are filtered out)
`POST /v1/assistants/chatConversations/{id}/leave`	Leave a shared conversation
`GET /v1/assistants/chatConversations/{id}/state`	Read conversation state
`POST /v1/assistants/chatConversations/{id}/stop`	Stop an in-flight response

Collaborators are emailed when they are added, appear as an avatar stack on shared chats in the sidebar, and are included in the conversation listing.

Conversation content that was previously visible only to its author can now be shared by that author. If your compliance posture assumed conversations were single-user, review it.

Chat experience changes

Change	Class	Detail
`Cmd`/`Ctrl` + `K`	Additive	Opens the Ask box globally in AI Mode
Persona-aware suggestions	Additive	Landing-page prompt suggestions follow the active persona
Redesigned prompt input	Behavioural	Gradient styling, chips and inline suggestions
Tool arguments surfaced	Additive	Per-call tool arguments are shown on chat messages
Rich markdown	Additive	Cards, pills and inline SQL render in responses. AI Automation entity links route correctly.
Native profile page	Additive	A profile page inside AI Mode
Artifact preview & scroll-to-bottom	Additive	Restored and added respectively
History trimming	Behavioural	Large tables and charts are trimmed out of conversation history sent to the model, so very long sessions stay responsive

Query Runner

Change	Class	Detail
Results stream to object storage	Behavioural	Large result sets are written to object storage and the UI is driven by websocket rather than held in memory
Azure result storage backend	Additive	In addition to the existing backend
`GET /v1/collate/apps/queryRunner/health/storage`	Additive	Storage health probe
`GET\|POST /v1/collate/apps/queryRunner/results/{workflowId}`	Additive	Fetch and publish results by workflow
`GET /v1/queryRunner/configs/defaults`	Additive	Read the effective defaults
`maxResultSize` cap raised to 10,000	Behavioural	Previously lower
`executionLimit` surfaced	Additive	The guardrail `LIMIT` applied to a query is returned on the response, so users can see their result was capped

Object-storage streaming means Query Runner now needs storage configured. Check the storage health endpoint after upgrading.

Natural-language query

Breaking. NLQ provider configuration moves into the unified llmConfiguration block. See Platform & Security.

Change	Detail
Provider abstraction	NLQ runs through a client abstraction with OpenAI and Google Gemini clients in addition to the existing provider
Filter extraction	Routed through the unified LLM configuration. Test suite and test case entity types are extracted.
Result shaping	Query-level collapse and result shaping for multi-chunk hybrid ranking
`GET /v1/hybrid/nlq/status`	New status endpoint

Collate desktop app

Breaking. The desktop app refuses to save an instance URL whose server is below 2.0.0.

Saving a Collate instance URL in desktop Settings probes GET /api/v1/system/version and, if the version parses below 2.0.0, reopens the dialog with an inline alert instead of saving. It fails open on an unparseable version. The Electron shell is now a thin wrapper, and title-bar back/forward navigation buttons are added.

Upgrade your Collate instances before rolling out the 2.0 desktop client, or users will be unable to point the app at them.

Slack and Microsoft Teams

Breaking. The Collate AI Slack app moves from assistant mode to agent mode, and now requires file-attachment permissions so it can deliver AI Analytics reports.

Reinstall the Slack app from the updated manifest and re-approve the requested scopes. Without the file permission, report delivery to Slack fails.

Slack rendering also moves to template-driven Block Kit, with numeric alignment, a PII callout, profile statistics and CDN-hosted service logos. Chart image processing outlasts Slack’s processing window rather than falling back to a degraded payload, and the app answers top-level DMs.

Response language is pinned from the profile

Behavioural. Collate AI’s response language is now taken from the X-OpenMetadata-Language header (set from the Language tab of profile settings) instead of being inferred from the conversation.

A user whose profile language differs from the language they type in will now get replies in their profile language. Tell users where the setting lives before they report it as a bug.

MCP

Change	Class	Detail
Tool annotations	Additive	Collate tools are annotated with title, read-only, destructive and open-world hints
Size budget	Behavioural	Collate tools route through the shared response size-budget net, with null-guarded error messages
Usage tracking	Additive	MCP tool-call usage is tracked and rolled up daily into telemetry

See API & Schema for the platform-level MCP contract changes, which also apply.

Collate 2.0

Breaking Changes by Component

Breaking Changes - Collate AI & Credits | Official Documentation

Collate AI & Credits

AskCollate is renamed to Collate AI

Credits are now weighted

Runs that cannot finish are refused

Billing is redesigned

Conversations become collaborative

Chat experience changes

Query Runner

Natural-language query

Collate desktop app

Slack and Microsoft Teams

Response language is pinned from the profile

MCP

​Collate AI & Credits

​AskCollate is renamed to Collate AI

​Credits are now weighted

​Runs that cannot finish are refused

​Billing is redesigned

​Conversations become collaborative

​Chat experience changes

​Query Runner

​Natural-language query

​Collate desktop app

​Slack and Microsoft Teams

​Response language is pinned from the profile

​MCP

Collate AI & Credits

AskCollate is renamed to Collate AI

Credits are now weighted

Runs that cannot finish are refused

Billing is redesigned

Conversations become collaborative

Chat experience changes

Query Runner

Natural-language query

Collate desktop app

Slack and Microsoft Teams

Response language is pinned from the profile

MCP