Ask questions across your Consilio matters, workspaces, and documents without leaving the chat.
The Aurora MCP server is Consilio's hosted bridge between an AI assistant and the eDiscovery engagements your team already runs on the Aurora platform. It answers questions about matters, workspaces, documents, and support tickets over a single Streamable HTTP endpoint at https://mcp.ai.consilio.com, and every response is filtered through the entitlements of whoever signed in. Nothing is written back — this is a read path by design.
Legal teams lose more hours to navigation than to analysis: finding the right alpha code, remembering which workspace holds the production set, chasing down a document someone referenced on a status call. Aurora collapses that into conversation. You ask for the matters tied to a client, the assistant resolves them, and you keep drilling — into workspaces, then documents — without switching tabs. Results carry source URLs, so anything the assistant asserts can be opened and verified in the Aurora UI rather than taken on faith.
This connector earns its keep when the bottleneck is recall across a large matter portfolio rather than raw document review. A case team preparing for a meet-and-confer, a project manager reconciling workspace status across a dozen engagements, or a partner asking what was produced on a given matter and when — all of those get answered in seconds. If you need bulk export, tagging, coding decisions, or anything that mutates the record, you still go to the Aurora application; the connector deliberately refuses to write.
Connection uses OAuth 2.1 with PKCE and RFC 9728 protected-resource discovery, so most clients need nothing beyond the URL. On first connect your client opens a browser, you authenticate with Consilio Identity (the same SSO you use elsewhere), and the token is reused afterwards. Claude Code, Claude Desktop, Cursor, VS Code, and GitHub Copilot all handle that flow natively.
This is a customer-gated connector. Without an active Consilio account and at least one entitled matter, the endpoint returns nothing useful — an unauthenticated request is rejected outright with a missing-JWT error. Consilio states that prompts and tool outputs are not retained and only short operational metrics are kept, which matters when you are running privilege-sensitive queries, but confirm that against your own engagement terms before pointing an assistant at live matters. Availability is also rolling out progressively rather than being switched on for every tenant at once, so check with your Consilio contact before planning workflows around it.
Pull the engagements tied to a client and summarise their current workspace layout minutes before a status meeting.
Ask a factual question across documents in an active matter and get an answer with links back to the underlying records.
Sweep across several engagements to spot workspaces or tickets that have gone stale without opening each one by hand.
| Tool | What it does |
|---|---|
| aurora_search_matters | Find Consilio engagements by client name, alpha code, or a natural-language description of the dispute. |
| aurora_search_workspaces | List the workspaces that sit under a matter so you can see where review or production work is organised. |
| aurora_search_documents | Search documents and their metadata inside an entitled matter and return citable source links. |
| aurora_search_tickets | Look up Consilio project and support tickets associated with an engagement. |
claude mcp add aurora --transport http https://mcp.ai.consilio.comAn active Consilio account with Aurora entitlements; the connector is read-only and authenticates through Consilio Identity SSO.
A knowledge graph your assistant keeps between sessions — entities, relations and observations that persist.
Kill hallucinated APIs — version-accurate, up-to-date library documentation injected straight into context.
Your workspace, on speaking terms with AI — search, read and write Notion pages and databases.
A structured scratchpad for hard problems — stepwise reasoning with revisions, branches and visible logic.
Symbol-level code navigation, refactoring and memory for coding agents — the IDE brain your assistant has been missing.
Chat with your second brain — search, read and write vault notes through the Local REST API.