3. OpenAPI Integration

Overview

Liman can automatically generate tools from OpenAPI specifications. This means your existing REST APIs become agent tools without writing custom code.

This builds on the Adding Custom Tools guide.

Step 1: Create a Test API

First, install FastAPI:

uv add "fastapi[standard]"

Create server.py with a simple user management API:

server.py

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel

app = FastAPI(title="User Management API", version="1.0.0")

class User(BaseModel):
    id: str
    name: str
    email: str

# Mock data
users = {
    "1": {"id": "1", "name": "Alice", "email": "alice@example.com"},
    "2": {"id": "2", "name": "Bob", "email": "bob@example.com"},
}

@app.get("/users/{user_name}", response_model=User, operation_id="get_user")
def get_user(user_name: str):
    """Get a user by their name"""
    for user in users.values():
        if user["name"].lower() == user_name.lower():
            return User.model_validate(user)

    raise HTTPException(status_code=404, detail=f"User {user_name} not found")

Step 2: Start Your API

Start the server:

fastapi run server.py

Visit http://localhost:8000/docs to see the auto-generated OpenAPI spec.

Step 3: Install `liman-openapi` Package

uv add liman-openapi

Step 4: Create ToolNodes from OpenAPI

Update your main.py:

main.py

from liman.agent import Agent
from liman_openapi import create_tool_nodes, load_openapi  
...

async def main():
    agent = Agent("./specs", start_node="LLMNode/chat", llm=get_llm())
    openapi = load_openapi("http://localhost:8000/openapi.json")  
    create_tool_nodes(openapi, agent.registry, base_url="http://localhost:8000")  

    ...


if __name__ == "__main__":
    asyncio.run(main())

Step 5: Add Tool to LLMNode

Update specs/chat.yaml:

specs/chat.yaml

kind: LLMNode
name: chat
prompts:
  system:
    en: |
      You are a helpful assistant.
      Always be polite and provide clear answers.
tools:
  - get_user_by_name
  - OpenAPI__get_user

Step 6: Test the Integration

Start the API server: fastapi run server.py (we already did this above)

Run the agent:

bash

python main.py

# for detailed output, you can enable debug mode:
# LIMAN_DEBUG=1 python main.py

Try these examples:

You: Who is Alice?
Agent: Alice is a user with the email address alice@example.com.

You: What emails do Bob and Pavel have?
Agent: Bob's email is bob@example.com. However, it seems there is no user named Pavel found in the system.

It's the same as in the previous example, but now we're calling your server!

How Auto-Generation Works

When you call:

main.py

openapi = load_openapi("http://localhost:8000/openapi.json")
create_tool_nodes(openapi, agent.registry, base_url="http://localhost:8000")

Liman:

Fetches OpenAPI spec from your running server
Generates ToolNode definitions for each endpoint
Creates HTTP request functions automatically
Registers tools with names like OpenAPI__get_user

Your OpenAPI endpoint becomes a callable tool without any manual work!

Next Steps

Try the python/samples/simple_openapi sample to see this integration in action with more complex scenarios.

3. OpenAPI Integration

Behind the scene

On this page