mcpforge
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@mcpforgelist files in the current directory"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
mcpforge
Ship an MCP server in 5 lines of Python.
Before After
────── ─────
~200 lines of @serve
JSON-RPC stdio class MarketTools:
boilerplate, schema @tool
generation, error def latest_price(self, symbol: str) -> float: ...
handling, lifecycle @tool
management. def search(self, query: str) -> list[dict]: ...
$ python -m mcpforge run market:MarketTools
✓ MCP server running on stdioWhy mcpforge
The Model Context Protocol (MCP) is the open standard Claude, Cursor, and an exploding ecosystem of AI tools use to call external functions. It's powerful — and writing a server is currently miserable.
You write the same JSON-RPC framing, schema generation, dispatch loop, and error handling on every project. mcpforge erases all of it. Annotate your class. Type-hint your methods. You're done.
from mcpforge import serve, tool
@serve(name="market_tools", version="0.1.0")
class MarketTools:
"""Market data tools for AI agents."""
@tool(description="Get the latest price for a symbol")
def latest_price(self, symbol: str) -> float:
return fetch_price(symbol)
@tool
def search(self, query: str, limit: int = 10) -> list[dict]:
"""Search ticker symbols matching `query`."""
return run_search(query, limit)That's a complete, spec-compliant MCP server. Type hints become JSON Schema. Docstrings become tool descriptions. Returns are auto-serialized.
60-second quickstart
pip install mcpforge# hello.py
from mcpforge import serve, tool
@serve(name="hello", version="0.1.0")
class HelloTools:
@tool
def greet(self, name: str = "world") -> str:
"""Say hello to someone."""
return f"Hello, {name}!"python -m mcpforge run hello:HelloToolsYou now have a working MCP server speaking JSON-RPC 2.0 over stdio.
How it works
@servetags a class as an MCP server (name, version, capabilities).@toolregisters methods as MCP tools.@resource(uri=...)registers methods as MCP resources.mcpforge introspects each method's signature and produces a JSON Schema — primitives,
list[T],dict[str, T],Literal[...],Optional[T], dataclasses, and Pydantic models all just work.python -m mcpforge run mod:Classboots the stdio loop, handlesinitialize/tools/list/tools/call/resources/list/resources/read, and reports JSON-RPC errors with proper codes.
No external MCP SDK. Just stdlib json + pydantic for schema niceties.
Built-in servers
Two batteries-included servers you can drop in today:
# Filesystem tools (sandboxed to a root directory)
python -m mcpforge run mcpforge.builtin.filesystem:FilesystemTools
# HTTP fetch tools
python -m mcpforge run mcpforge.builtin.http:HttpToolsFilesystemTools exposes list_dir, read_file, search — sandboxed with
path traversal checks. HttpTools exposes fetch_url with size limits.
Plug into Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"market": {
"command": "python",
"args": ["-m", "mcpforge", "run", "market:MarketTools"],
"cwd": "/path/to/your/project"
}
}
}Restart Claude Desktop. Your tools appear in the conversation.
Plug into Cursor
~/.cursor/mcp.json:
{
"mcpServers": {
"market": {
"command": "python",
"args": ["-m", "mcpforge", "run", "market:MarketTools"]
}
}
}Comparison
Feature | mcpforge |
| FastMCP | Hand-rolled |
Single decorator | Yes | No | Yes | No |
Auto JSON Schema from types | Yes | No | Yes | No |
Pydantic v2 support | Yes | Yes | Yes | No |
Zero required deps beyond pyd | Yes | No | No | N/A |
Built-in fs / http servers | Yes | No | No | No |
Lines for hello world | ~5 | ~40 | ~10 | ~200 |
Inspect the wire format
python -m mcpforge inspect market:MarketToolsPrints the exact tools/list payload your clients will see.
Roadmap
Tools (call, list)
Resources (read, list)
Type-hint -> JSON Schema (Pydantic v2, dataclasses, Literal, Optional)
Built-in filesystem and http servers
Resource subscriptions
Prompts capability
Sampling capability
WebSocket / HTTP-SSE transport (optional
[http]extra)Async tools (
async def)OTel tracing hooks
License
MIT — see LICENSE.
Author: thechifura. Sibling to quantflow and strategos.
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/vigilancetrent/mcpforge'
If you have feedback or need assistance with the MCP directory API, please join our Discord server