
Bolt Claw MCP server
Which MCP servers are Bolt Claw's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.
read 2026-11-05
- Registry nameai.boltclaw/docsverified in the MCP registry
The registry name sits under boltclaw.ai. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.
- Stateanswerschecked 2026-11-05
- Tools2
- Server calls itselfboltclaw-docs
- Server version1.22.0
- Protocol version2025-06-18
On 2026-11-05 the server answered initialize and tools/list without a sign-in. First answer in 161 ms.
Should an agent connect it
- On Nov 19, 2026 we called its read-only tools following a fixed scenario, and 1 of 3 steps returned what the check expects.
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.
Measured daily
- INITE MCP score100 of 100answers 100 · definitions 100 · context 100 · connects 100
- Checked by calling it1 of 3 steps passedlatest run Nov 19, 2026, read-only calls only
- Answered100% of days18 of 18 daily checks
- Median first answer216 ms
- Space in an agent's context≈ 234 tokensthe definitions of all its tools (2), which an agent carries on every turn
- Tool definitionsA · 100 of 100
- Typed results0 of 2 toolsdeclared outputSchema: results come typed, read once
- keyword-search, version 1, Nov 19, 2026: 1 of 1 steps passed. See the evidence
- search-then-fetch, version 1, Nov 12, 2026: 0 of 2 steps passed. See the evidence
- Every tool passes all six checks.
Worth a look
None of the five warning signs in these definitions.
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 →
Grade A, 98 of 100, checked 2026-11-22 · 1 failed
searchverified Nov 19, 2026Find documentation pages about a subject. Returns an id, a title, a URL and a matching line for each hit, best first. Pass the id to `fetch` to read the page.
fetchseen, not called yetRead one documentation page in full, as markdown. The id is what `search` returns, which is the page's path without /docs/ - for example `authentication` or `api/pages`.
How to connect
Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:
{
"mcpServers": {
"docs": {
"type": "http",
"url": "https://boltclaw.ai/docs/mcp-server"
}
}
}Add it
Claude Code, in a terminal
claude mcp add --transport http boltclaw https://boltclaw.ai/docs/mcp-serverclaude.ai, ChatGPT and other clients: add https://boltclaw.ai/docs/mcp-server as a custom connector in the settings.
OpenAPI spec
170 paths, read 2026-11-05. Open the spec ↗
Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.
Before we count this visit
Analytics and ad measurement tell us which pages and which of our ads do their job: Google Analytics and Google Ads, the Meta pixel, Yandex Metrica, and one first-party id of our own. Nothing is stored until you answer, and declining changes nothing about the site.