verdoc-mcp
Provides coding agents with repository context through tools that generate AGENTS.md files or structured JSON summaries, quoting build/test commands from the repo's manifests for a small fee without an API key.
README
verdoc-mcp
Give your coding agent the context of a repository it has never seen, for two cents, without an account or an API key.
An agent dropped into an unfamiliar codebase has to read it before it can work
in it - and that costs tens of thousands of tokens every time. verdoc-mcp
exposes two tools that answer the same questions directly:
| tool | what you get | price |
|---|---|---|
verdoc_agents_md |
an AGENTS.md ready to commit |
$0.02 |
verdoc_scan |
the same analysis as structured JSON | $0.01 |
Build and test commands are quoted from the repository's own manifests, with
the file each one came from named. Every path is checked against the git tree.
Anything ambiguous is marked unverified rather than guessed.
No language model runs on the server. The same commit always returns the same answer.
Install
Not on npm yet - install straight from this repository:
{
"mcpServers": {
"verdoc": {
"command": "npx",
"args": ["-y", "github:verdochello/verdoc-mcp"],
"env": { "EVM_PRIVATE_KEY": "0x..." }
}
}
}
Or clone it and point at the file directly:
git clone https://github.com/verdochello/verdoc-mcp && cd verdoc-mcp && npm install
{
"mcpServers": {
"verdoc": {
"command": "node",
"args": ["/absolute/path/to/verdoc-mcp/index.mjs"],
"env": { "EVM_PRIVATE_KEY": "0x..." }
}
}
}
Works with Claude Desktop, Claude Code, Cursor, or any MCP client.
Without a key it still runs. The tools report what they would cost and what they would return, and buy nothing. That is the intended way to try it.
Paying
The wallet needs USDC on Base, Polygon or Arbitrum. It does not need ETH - x402 payments are signed off-chain and the facilitator pays the gas.
| variable | default | meaning |
|---|---|---|
EVM_PRIVATE_KEY |
- | wallet that pays. Omit to run in quote-only mode. |
VERDOC_MAX_SPEND_USD |
0.05 |
hard per-call ceiling. Nothing above it is ever signed. |
VERDOC_ORIGIN |
https://verdoc.dev |
override the endpoint (testing). |
About that private key
You are handing a program the ability to spend your money, so here is exactly what this one does and does not do.
- The key is read from the environment. It is never logged, never written to disk, and never sent anywhere. Error messages are scrubbed of anything that looks like a key before being returned.
- The payee and the chain are pinned in the source. Before signing anything, the client reads the live payment offer and checks it. If Verdoc ever asks to be paid at a different address, or on a chain not in the list, this refuses and tells you instead of paying. That means a compromise of our own domain or DNS cannot redirect your funds.
- Every call is capped by
VERDOC_MAX_SPEND_USD, checked before a signature exists.
Use a wallet you funded for this purpose. That is good practice with any paying agent tool, including this one.
What it will refuse to do
These are tested in selftest.mjs, and the refusals are the
part worth reading:
- pay an address other than the pinned one - even if the live endpoint asks
- pay on a chain outside the pinned set, including testnets
- pay more than your ceiling
- treat a short or empty 200 response as a real answer
Limits
Public GitHub repositories only. Repositories over 20,000 tracked files are rejected: GitHub truncates the git tree above that, and every claim this service makes depends on having the complete tree.
Licence
MIT. The service it calls is at verdoc.dev - OpenAPI | x402 discovery
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.