MCP-Typescribe - an MCP Server providing LLMs API information
@yWorks
About MCP-Typescribe - an MCP Server providing LLMs API information
An MCP server implementation enabling LLMs to work with new APIs and frameworks
Config
Add this server to your MCP-compatible client using the configuration below.
{
"mcpServers": {
"mcp-typescribe": {
"command": "npx",
"args": [
"typedoc",
"--json",
"docs/api.json",
"--entryPointStrategy",
"expand",
"path/to/your/typescript/files"
]
}
}
}Tools
No tools detected
Fetch the live tool list by running this server in a temporary sandbox using the button above.
Overview
What is MCP-Typescribe?
MCP-Typescribe is an open-source MCP server that provides large language models (LLMs) with real-time access to TypeScript API documentation. It loads TypeDoc-generated JSON and exposes query endpoints so AI coding assistants can search for symbols, get detailed information, and understand relationships within unfamiliar TypeScript APIs.
How to use MCP-Typescribe?
You generate TypeDoc JSON for your TypeScript API, build the project with npm run build, then run the server using npx mcp-typescribe@latest run-server <path-to-api.json>. AI agents such as Cline in VS Code connect to the server by specifying it in their MCP settings (e.g., cline_mcp_settings.json). The server exposes tools like search_symbols and get_symbol_details for agents to query.
Key features of MCP-Typescribe
- Loads and indexes TypeDoc JSON documentation
- Search symbols by name with optional kind filtering
- Get detailed information about any symbol
- List methods and properties of classes or interfaces
- Find implementations of interfaces or subclasses
- Show inheritance relationships between types
Use cases of MCP-Typescribe
- Enable LLM-based coding assistants to use new or proprietary TypeScript SDKs
- Let AI agents explore and understand unfamiliar APIs without retraining
- Query internal API documentation dynamically rather than putting all docs into context
- Help developers onboard faster by letting LLMs discover API symbols and usage patterns
FAQ from MCP-Typescribe
What does MCP-Typescribe do that alternatives do not?
It converts TypeDoc JSON into a machine-readable format served via MCP, so LLMs can query specific API details on demand instead of relying on static training data or full documentation files.
What are the runtime requirements?
Node.js and npm. The server runs locally and connects to AI agents via the Model Context Protocol.
Where does the documentation data live?
You provide a local TypeDoc-generated JSON file (e.g., docs/api.json). The server loads and indexes this file at startup.
What transports and authentication does it use?
The README only describes a local stdio-based MCP configuration; no network transports or authentication methods are mentioned.
Is this project still actively developed?
Public development is currently suspended because no active community formed. The team is working on a separate, follow-up project for the yFiles API.
Frequently asked questions
What does MCP-Typescribe do that alternatives do not?
It converts TypeDoc JSON into a machine-readable format served via MCP, so LLMs can query specific API details on demand instead of relying on static training data or full documentation files.
What are the runtime requirements?
Node.js and npm. The server runs locally and connects to AI agents via the Model Context Protocol.
Where does the documentation data live?
You provide a local TypeDoc-generated JSON file (e.g., `docs/api.json`). The server loads and indexes this file at startup.
What transports and authentication does it use?
The README only describes a local stdio-based MCP configuration; no network transports or authentication methods are mentioned.
Is this project still actively developed?
Public development is currently suspended because no active community formed. The team is working on a separate, follow-up project for the yFiles API.
Basic information
More Developer Tools MCP servers
ship.page
BitgateThrow HTML at an API, get an unguessable URL back. Zero-config remote MCP server — your agent deploys pages, sites and CI reports in one call. Free, no signup needed.
directree
coburn2716directree MCP server — query the honest software directory from any AI assistant
Hermoso
HermosoHermoso is a remote MCP server that runs a full marketing operation from your agent: research competitor ads, generate image and video ads, publish to social channels, and build and manage paid campaigns. 718 tools (306
LocalCan
LocalCanGives AI agents public URLs (tunnels) for localhost, live HTTP traffic inspection, snapshot publishing, and access control. Part of LocalCan, the ngrok alternative for Mac, Windows and Linux. Free plan.
Raccha AI
Shree MandadiMCP-first toolbox for agents: namespaced key-value storage, a FIFO queue, auth, and a handful of stateless utility tools (JWT decode, hashing, cert inspection, CIDR math). Free while in early access. This repository is a
Comments