Dossier MCP Server

Dossier — Google Gemini Deep Research as an MCP server. Durable background runs, spend control, context-safe report reading, citation verification, and private-corpus grounding. Exposed over MCP by the dossier mcp server, that capability becomes something an assistant can invoke while it works, not something you go and do afterwards.

What it actually does

Not a web search. A method it can't skip, citations you can check, and no API key required to start.

Its toolset

Everything the assistant can do here goes through one of these:

• Mid — run progress is not available on Gemini.** Its API buffers it; a 7.1-minute run reported nothing until it finished. The plumbing is in place if that
• Private — corpus grounding is Gemini only.** OpenAI vector stores and xAI collections are real features and aren't wired up here yet, so both declare no corpus
• Picks — technical archetype, max tier
• Reads — Project docs, GitHub issues, benchmark posts, release notes
• Then — Call it again with the run handle and it checks the returned table against what you asked for
• Windows — 24h, 7d, 30d, 90d, 1y, 5y, all. Defaults to 30 days here and a year everywhere else
• Costs — A full run on each backend. Worth it when a number is load-bearing
• Start — One search task per source class, each with the query dialect that index expects
• Draft — The registry freezes. No source can be added after that, including by you
• Submit — Every cited URL is checked against the frozen registry, and the draft is refused if it cites anything you never gathered
• xAI — model's choice, capped

Adding it to your client

Being a remote server, there is no local install. You register the endpoint with your client, authorise it once, and the tools appear.

Configuration

You will need 4 environment variables: GEMINI_API_KEY, PERPLEXITY_API_KEY, OPENAI_API_KEY, XAI_API_KEY. The server will not start without them, which is usually why the tools fail to appear on a first run. Keep credentials in your client's env block or a secrets manager rather than in a file you might commit.

When to reach for it

Plenty of developer tooling servers cover similar ground. The differences that matter in practice are scope of access and how much setup stands between you and a working tool call. Dossier's toolset — Mid, Private, Picks and 8 more — is a fair guide to whether it matches your workflow. It is maintained by lukerhodes; worth a glance at recent repository activity before you build anything load-bearing on it.

SyncDev reviews every entry in this directory against the project's own documentation before publishing, and revisits them as servers change.

Caveats

• Your data travels to the provider's service, so the usual questions apply about what you send and what they retain.
• With 11 tools registered it takes up a noticeable share of the context window; turn it off in projects that never touch Dossier.
• Missing credentials fail quietly in some clients — if no tools show up, check the environment block first.
• MCP clients confirm each tool call by default. Leave that on until you have watched what the dossier mcp server does with a few real requests.