Fixie AI

Build conversational AI agents with integrated tool use and memory using Fixie AI. This Action Pack guides you through setting up your agent, defining its capabilities, and deploying it to interact intelligently with users.

beginner30 min5 steps

The play

Install Fixie CLI & Authenticate
Install the Fixie Command Line Interface (CLI) and authenticate your account to manage agents. This enables local development and deployment.
Define Your First Agent
Create a new Python file to define your basic Fixie agent. Implement the `handle_message` method to process incoming user messages and generate responses.
Implement Tool Use
Integrate external functions or APIs as 'tools' your agent can call. Define a tool and register it within your agent to enable actions beyond simple text generation.
Add Memory to Your Agent
Configure your agent to remember past interactions or user-specific data using Fixie's memory capabilities. This allows for more personalized and context-aware conversations.
Deploy and Test Your Agent
Publish your agent to the Fixie platform using the CLI. Test its functionality, conversational flow, and tool integrations in a live environment.

Starter code

from fixieai.agents import Agent
from fixieai.agents.models import Message

class EchoAgent(Agent):
    def handle_message(self, message: Message):
        return f"Echo: {message.text}"

# To run this agent locally:
# 1. Save it as 'echo_agent.py'
# 2. Run in your terminal: fixie agent run echo_agent.py EchoAgent

Source

Articlefixie.ai