# embedded-docs

Hands your coding agent verbatim documentation snippets with section and page citations. Useful when a register value has to be exact.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 26
- Author: ByteAsk
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/ByteAsk/ByteAsk-Embedded-MCP
- Updated: 2026-06-20
- HTML: https://affpapa.org/en/nevibecode/mcp/embedded-docs

## What it does

Queries go in as plain language or an exact identifier. What comes back is ranked evidence from the corpus: document title, section and page reference, the source text itself, and a `result_id`. A miss returns "no confident match" instead of a made-up register value. The `get_context` tool expands a hit into its full section, and `request_document` logs a request for a document that is missing from the corpus.

Three tools, two transports. Local agents talk over `stdio`; the hosted endpoint speaks Streamable HTTP at `https://mcp.byteask.ai/mcp`. The retrieval engine and the licensed corpus live outside this repository, behind a two-method `SearchBackend` interface. A bundled `SampleBackend` with a few illustrative records lets `search_docs` answer right away, no API keys involved. Point the server at your own engine with `BYTEASK_BACKEND` set to `module:callable`.

Running it locally takes Python 3.10 or newer and `uv`. Settings come from environment variables and `.env`: transport, HTTP host and port, bearer token, allowed `Host` headers. Leaving `MCP_HTTP_AUTH_TOKEN` empty drops authentication from the HTTP server, which is fine on your machine and a bad idea anywhere reachable from the network.

## Requirements
- Needs: Python and uv (commands run via uvx)
- Remote server: `https://mcp.byteask.ai/mcp`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http embedded-docs https://mcp.byteask.ai/mcp
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add embedded-docs --url https://mcp.byteask.ai/mcp
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "embedded-docs": {
            "url": "https://mcp.byteask.ai/mcp"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "embedded-docs": {
            "type": "http",
            "url": "https://mcp.byteask.ai/mcp"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http embedded-docs https://mcp.byteask.ai/mcp
```

## When to use it
- confirm a register offset before touching driver code
- look up the Modbus function code for writing multiple registers
- check an IEEE clause rather than trust recall
- pull the exact SCPI command syntax
- ask for a missing datasheet to be added to the corpus

## Example prompts
- Search the docs for the Modbus function code that writes multiple registers
- Check the datasheet for the ETH_DMATDLAR register offset and cite the page
- Expand the full section for this `result_id`

## Similar
- [REA](https://affpapa.org/en/nevibecode/mcp/rea.md) — MCP server · ★ 386989
- [GitHub](https://affpapa.org/en/nevibecode/mcp/github.md) — MCP server · ★ 33918
- [Netdata](https://affpapa.org/en/nevibecode/mcp/netdata.md) — MCP server · ★ 81115
- [browser-use](https://affpapa.org/en/nevibecode/mcp/browser-use.md) — MCP server · ★ 119160
- [Codebase Memory](https://affpapa.org/en/nevibecode/mcp/codebase-memory.md) — MCP server · ★ 47392
- [mcp-server](https://affpapa.org/en/nevibecode/mcp/mcp-server.md) — MCP server · ★ 43872

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md