agent-service-toolkit: A Comprehensive Toolkit for AI Agent Services with LangGraph

Introduction

The agent-service-toolkit by JoshuaC215 offers a complete solution for running AI agent services. Built with LangGraph, FastAPI, and Streamlit, it provides everything needed from agent definition to a user-friendly chat interface. This project serves as an excellent template for developers looking to quickly build and deploy their own agents using the LangGraph framework, demonstrating a full, robust setup.

Key components include a LangGraph agent, a FastAPI service to serve it, a client for interaction, and a Streamlit application for a chat interface. Data structures and settings are meticulously built with Pydantic, ensuring reliability and ease of use.

Installation

Getting started with agent-service-toolkit is straightforward, with options for both Python virtual environments and Docker.

Quickstart with Python

Set up environment variables: Create a .env file in the root directory with at least one LLM API key (e.g., OPENAI_API_KEY=your_openai_api_key).

Install dependencies: Use uv (recommended) or pip.

curl -LsSf https://astral.sh/uv/0.7.19/install.sh | sh
uv sync --frozen
source .venv/bin/activate

Run the service: In one terminal:
```
python src/run_service.py
```

Run the Streamlit app: In another terminal:

source .venv/bin/activate
streamlit run src/streamlit_app.py

Quickstart with Docker

Set up environment variables: Create a .env file with your OPENAI_API_KEY.
```
echo 'OPENAI_API_KEY=your_openai_api_key' >> .env
```
Launch services: Ensure Docker and Docker Compose (>= v2.23.0) are installed.
```
docker compose watch
```
This command will automatically start the PostgreSQL database, the FastAPI agent service, and the Streamlit app. Services will update automatically on code changes.

Examples

The repository includes a generic src/client/client.AgentClient for interacting with the agent service. Here's a quick example of how to use it:

from client import AgentClient
client = AgentClient()

response = client.invoke("Tell me a brief joke?")
response.pretty_print()
# Expected Output:
# ================================== Ai Message ==================================
#
# A man walked into a library and asked the librarian, "Do you have any books on Pavlov's dogs and Schrödinger's cat?"
# The librarian replied, "It rings a bell, but I'm not sure if it's here or not."

Why Use It

agent-service-toolkit stands out due to its comprehensive feature set and robust architecture:

LangGraph Agent and Latest Features: Implements LangGraph v1.0 features, including human-in-the-loop with interrupt(), flow control with Command, long-term memory with Store, and langgraph-supervisor.
FastAPI Service: Provides both streaming and non-streaming endpoints for serving agents efficiently.
Advanced Streaming: Features a novel approach supporting both token-based and message-based streaming.
Streamlit Interface: Offers a user-friendly chat interface with voice input and output capabilities.
Multiple Agent Support: Allows running and calling multiple agents by URL path, with available agents and models described in /info.
Asynchronous Design: Utilizes async/await for efficient handling of concurrent requests.
Docker Support: Includes Dockerfiles and a docker compose file for easy development and deployment.
Testing: Comes with robust unit and integration tests for the entire repository.

agent-service-toolkit: A Comprehensive Toolkit for AI Agent Services with LangGraph

Summary

Repository Info

Tags