Skip to main content
Agentor is an open-source framework that makes it easy to build AI Agents, distributed multi-agent systems, MCP Servers or ANY LLM application with secure integrations across email, calendars, CRMs, and more.

​
Start with four steps

1

Install

pip install agentor
2

Build an Agent

Once installed, you can start building your first Agent with just a few lines of code.
from agentor.tools import WeatherAPI
from agentor import Agentor

agent = Agentor(
    name="Weather Agent",
    model="gpt-5-mini",
    llm_api_key="sk-xxx",
    tools=[WeatherAPI()]
)
result = agent.run("What is the weather in London?")

{
  "response": "The current weather in London is 15°C with partly cloudy skies. Wind speed is 12 km/h from the northwest."
}
3

Run the Agent as a server

Agents can be run as a server to be integrated into other applications such as UI applications, or other LLM applications.
from agentor.tools import WeatherAPI
from agentor import Agentor
  
agent = Agentor(name="Weather Agent", tools=[WeatherAPI()])
agent.serve() # Single line of code to serve the Agent
Deploy your Agent to cloud with a single command: celesto deploy
4

Query the Agent server

You can query your Agent server with a simple curl command.
curl -X 'POST' \
  'http://localhost:8000/chat' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "input": "What is the weather in London?"
}'

​
Troubleshooting

Common causes:
  • Missing or invalid API key
  • Network connectivity issues
  • Rate limiting from your LLM provider
Solution: Verify your API key is set correctly and check your provider’s rate limits.
Common causes:
  • Tool requires authentication or API keys
  • Invalid tool name or tool not available
Solution: Check the Tool Documentation for specific tool requirements.
Common causes:
  • Port 8000 is already in use by another application
Solution: Specify a different port: agent.serve(port=8001)
Common causes:
  • Agentor not installed correctly
  • Using wrong Python environment
Solution: Verify installation with pip show agentor and ensure you’re using the correct Python environment.

Still need help?

Join our Discord community for support, questions, and discussions with other developers.

​
Next steps

Build an Agent

Learn how to build an Agent with tools.

Build an MCP Server

Learn how to add capabilities to your Agent with MCP Servers.