Answer a question from the team's chat, wiki and tracker
No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.
What counts as done
An answer citing at least one linked source page, thread or issue the user can open, built from a full-text search, with no content from spaces the user cannot see.
Measured, for our own tests: at most five reads. No docs we read state these figures, so they don't count against a listing here.
How agents ask for it: what did we decide about the launch
, where is the onboarding doc
, summarise the status of project X
, what does the wiki say about deploys
Does the job, with a limit 8
It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.
| Listing | Why |
|---|---|
| Atlassian Rovo MCP Server C Atlassian Pty Ltd | searchConfluence (CQL), searchJiraIssuesUsingJql and a Beta Rovo search cover Jira, Confluence and connected apps under the user's own permissions. Cloud only: acc-19 used a third-party server for Data Center, and acc-31 measured 27,650 tokens of tools before the first call. source |
| Linear MCP C Linear Orbit, Inc. | The MCP finds issues, projects, comments and documents, and Linear search covers titles, descriptions and comments; it reaches only the tracker, not chat. The tool list isn't published, and acc-07 left the MCP for a CLI over token use. source |
| Mattermost C Mattermost, Inc. | POST /api/v4/posts/search runs a full-text search over posts the token's user can see, and GET /posts/{id}/thread reads a thread. It covers chat only, with no wiki or tracker, so answers drawn from docs or tickets need another tool. source |
| Microsoft Teams (Microsoft Graph) D Microsoft Corporation | The Microsoft Search API queries chatMessage bodies for the signed-in user, and messages carry a webUrl to cite. The listing is Teams chat only, so wiki or tracker answers need other Graph resources, and acc-23 never got the admin consent Graph needed. source |
| Notion MCP B Notion Labs, Inc. | notion-search runs content searches over pages and databases with filters, up to 50 results, and acc-13 found it full-text where the public API is title-only. Slack and Drive need paid AI plans, acc-12's legal team blocked it for lack of narrowing, and acc-33 got only a schema from fetch. source |
| Outline D General Outline, Inc. | documents.search searches document text with keywords and filters, limited to what the token can access, separate from documents.search_titles, and documents.info returns the Markdown body and url. It covers the wiki only, not chat or tickets. source |
| Slack MCP Server (official) B Slack Technologies, LLC (a Salesforce company) | The server searches messages and files, reads channels, threads and canvases, with private search on separate user-token scopes; acc-04 found answers good and faster than Glean. Only Marketplace or internal apps may connect (acc-06), and acc-01's standard connection exposed private messages. source |
| Zulip C Kandra Labs, Inc. | GET /messages with a narrow carrying search keywords, channel and topic operators runs a full-text search over messages the account can see. Zulip is chat only, so wiki or tracker answers need another tool. source |
What stops agents finishing
- public API searches titles only
- the deployment is Data Center or self-hosted and the hosted MCP does not reach it
- connection exposes private content an organisation forbids
Around it
Before: chat.user-question. After: productivity.write-doc, productivity.post-message.
Where this comes from
Weight 23 of 100 in Work & productivity: 7 of 36 accounts: Slack, Jira and Docs MCPs for status, Cursor beating Glean, Confluence Data Center search, Notion full-text search, a Notion wiki page whose body vanished
The same as data: /api/v1/use-cases/productivity.find-context.json, or the MCP tool get_use_case with id productivity.find-context. Other jobs in this category: /api/v1/niches?category=productivity.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.