Server facts
Grounded Docs MCP Server
Project
- Author
- arabold
- Repository
- github.com/arabold/docs-mcp-server
- Website
- grounded.tools
- License
- MIT
- Category
- Memory & Knowledge
Tech
- Runtime / language
- node
- Transports
- stdio
- Remote available
- No
- Authentication
- Not specified
Activity
- GitHub stars
- 1,812
- Last commit
- Data as of
Tags
- Documentation
- Knowledge Index
- Local Files
- Semantic Search
- Version Specific
Overview
Grounded Docs MCP Server gives coding assistants a documentation index tied to the libraries and versions used in a project. It fetches material from official documentation sites, GitHub, npm, PyPI, and local sources rather than relying only on a model's stored knowledge. Supported inputs extend beyond web pages and Markdown to PDFs, Office documents, notebooks, archives, source code, and structured data. Users can index documentation, search it through the command line, or retrieve an individual page as Markdown. A Web UI provides another way to add documentation, and a long-running MCP endpoint makes the index available to compatible assistants.
The README requires Node.js 22 or later for the npx workflow and also documents Docker deployment. Embeddings are optional: enabling an embedding provider adds semantic vector search, with OpenAI shown as one example and other providers referenced. Web crawling automatically looks for llms.txt links and prefers Markdown variants or Markdown content negotiation when supported by the source site. Hash-routed single-page documentation needs explicit hash preservation; that setting can trigger Playwright rendering because plain fetching cannot evaluate those routes. CLI output supports structured formats while keeping diagnostics separate from non-interactive stdout. The project is MIT licensed and documents deployment security and authentication topics separately from its local quick start.
Install & configure
Requirements
- Node.js 22 or later
- An MCP-compatible client for MCP access
npx @arabold/docs-mcp-server@latestThe README starts this long-running server, opens its Web UI at http://localhost:6280, and connects MCP clients to the local SSE URL. OPENAI_API_KEY is optional for OpenAI embeddings. CLI and Docker HTTP alternatives are documented.
Examples are translated from the published configuration. Supply your own credentials and verify requirements in the upstream README. Remote connectors are shown only where a supported configuration format is known.
Config file · macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
Use Settings → Developer → Edit Config when available. Do not assume a Linux configuration path; check the client documentation for your platform. Remote servers use the app’s connector flow when supported, not the local stdio file shown here. Merge this example with existing settings; do not overwrite unrelated servers.
A supported configuration example is not available for this client and transport. Check the client documentation for its connection flow; do not substitute an undocumented JSON format.
Read the Claude Desktop setup guideConfig file · Project: .cursor/mcp.json
User: ~/.cursor/mcp.json
The project path is relative to the workspace root. The tilde denotes your user home directory. Use Cursor’s MCP settings to confirm which configuration and scope your installed version is reading. Merge this example with existing settings; do not overwrite unrelated servers.
{
"mcpServers": {
"docs-mcp-server": {
"url": "http://localhost:6280/sse"
}
}
}Config file · Workspace: .vscode/mcp.json
For user scope, use the “MCP: Open User Configuration” command if available. Do not guess a user configuration path across operating systems, profiles, containers, or remote workspaces; check the client documentation. Merge this example with existing settings; do not overwrite unrelated servers.
A supported configuration example is not available for this client and transport. Check the client documentation for its connection flow; do not substitute an undocumented JSON format.
Read the VS Code setup guideConfig file · macOS/Linux: ~/.codeium/windsurf/mcp_config.json
Windows: %USERPROFILE%\.codeium\windsurf\mcp_config.json
Open the MCP configuration through Windsurf’s settings when possible to confirm the active file. Paths can differ in managed or future releases; check the client documentation if this location is not present. Merge this example with existing settings; do not overwrite unrelated servers.
{
"mcpServers": {
"docs-mcp-server": {
"serverUrl": "http://localhost:6280/sse"
}
}
}Config file · Project scope: .mcp.json at the repository root
The project file uses mcpServers. For user scope, use claude mcp add --scope user rather than assuming .mcp.json in your home directory is the user configuration. Check the client documentation for managed settings and storage details. Merge this example with existing settings; do not overwrite unrelated servers.
{
"mcpServers": {
"docs-mcp-server": {
"type": "sse",
"url": "http://localhost:6280/sse"
}
}
}Examples are translated from the published configuration. Supply your own credentials and verify requirements in the upstream README. Remote connectors are shown only where a supported configuration format is known.
Tools
Source: repository README. Tool descriptions are a documented snapshot, not a live capability check.
| Tool | What it does |
|---|---|
| scrape_docs | Index documentation, with an optional setting to preserve hash-based routes. |
Names, permissions, and defaults can change. Check the current upstream documentation before granting access.
Use cases
- Ground coding advice in documentation for the project's exact library version.
- Build a searchable index from local documents and repositories.
- Fetch a documentation page as Markdown for an agent or script.
- Index hash-routed documentation sites with explicit route preservation.
Strengths
- Supports diverse document, archive, code, and data formats.
- Offers CLI, Web UI, and MCP access to the documentation index.
- Embeddings are optional rather than required for startup.
- Automatically prefers available Markdown-oriented web sources.
Limitations
- Hash-routed sites require explicit preservation settings and browser rendering.
- Semantic vector search requires configuring an embedding model.
Frequently asked questions
Independent listing, not an affiliation or security review. Grant the least access needed and verify the upstream instructions before running code.