Corpusfetch MCP server

• Registry namecom.corpusfetch/corpusfetchverified in the MCP registry

The registry name sits under corpusfetch.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

• Stateanswerschecked 2026-11-02
• Tools4
• Server calls itselfcorpusfetch-mcp
• Server version0.1.0
• Protocol version2025-06-18

On 2026-11-02 the server answered initialize and tools/list without a sign-in. First answer in 252 ms.

• Its tools do every use case we checked: verify a warrant seal before fetching a record; route runtime exceptions to matching records via lookup.
• Checked on 2 of the 7 days it takes to judge reliability.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

• INITE MCP scoreafter a weeka score needs 7 daily checks: 2 so far
• What an agent can do with it2 of 2 use cases
• Answeredmeasuring2 of the 7 daily checks it takes to print a share
• Median first answer258 ms
• Space in an agent's context≈ 550 tokensthe definitions of all its tools (4), which an agent carries on every turn
• Tool definitionsB · 83 of 100
• Typed results0 of 4 toolsdeclared outputSchema: results come typed, read once

• Not saying whether it only reads or can destroy: 4 of 4

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

• corpusfetch_fetch_record

  Fetch the body of a CorpusFetch record. Format 'md' (default, ~1500 tokens) or 'py' (zero-prose stdlib slice, ~300 tokens). Prefer warrant-first: call corpusfetch_fetch_warrant before this.

• corpusfetch_fetch_warrant

  Fetch the ~220-token warrant (precondition seal) for a CorpusFetch record. Fetch this before the .md record. If the seal fails, do not fetch the record.

• corpusfetch_lookup_exception

  Route a Python exception class to candidate record addresses in the CorpusFetch pillar. Returns warrant_url, py_url, md_url, and fetch_when for each candidate. Fetch the warrant first (~220 tokens); if its seal fails, do not fetch the .md.

• corpusfetch_lookup_symptom

  Route a runtime symptom to candidate record addresses in the CorpusFetch pillar. Returns warrant_url, py_url, md_url, and fetch_when for each candidate. Fetch the warrant first (~220 tokens); if its seal fails, do not fetch the .md.