Getting Started

This guide takes you from zero to a running API in a few minutes.

0. Configure Environment

Create .env in your project root:

OPENAI_API_KEY=
GEMINI_API_KEY=
OPENROUTER_API_KEY=
DEFAULT_PROVIDER=openai

1. Create main.py

from agentapi import Agent, AgentApp

app = AgentApp()

agent = Agent(
    system_prompt="You are a helpful assistant",
    provider="openai",
)


@app.chat("/chat")
async def chat(message: str):
    return await agent.run(message)


@app.chat("/stream")
async def stream_chat(message: str):
    return agent.stream(message)

2. Run the Server

uvicorn main:app --reload

3. Test Endpoints

Chat

curl -X POST "http://127.0.0.1:8000/chat?message=Hello"

Stream

curl -N -X POST "http://127.0.0.1:8000/stream?message=Write%20a%20short%20poem"

4. Explore Interactive Docs

• Swagger UI: http://127.0.0.1:8000/docs
• ReDoc: http://127.0.0.1:8000/redoc

5. Use the CLI (Optional)

agentapi new myproject
cd myproject
uvicorn main:app --reload

The scaffold includes:

• main.py
• tools.py
• agents.py
• .env

6. Add Your First Tool

from agentapi import tool


@tool(
    description="Get the current date in ISO format.",
    context="Use when user asks for date/time.",
)
def get_date() -> str:
    from datetime import datetime
    return datetime.now().isoformat()

Attach it via Agent(..., tools=[get_date]).