Agent Functions Library

Why This Library?

Agent Functions Library is designed for the emerging world of AI-driven software development, where autonomous agents need to trigger and compose functions across different languages and platforms. It stands out by offering:

1. Native Agent Integration: Built specifically for AI agents, not adapted after the fact

   • Declarative function triggers that agents can easily understand and use
   • Built-in metadata and documentation that helps agents make decisions
   • Automatic validation and error handling suited for agent interactions

2. Cross-Language Support: Seamlessly work with both Python and JavaScript

   • Write functions in either language
   • Call Python functions from JavaScript and vice versa
   • Unified interface across languages

3. Cloud-Native & Serverless Ready

   • AWS Lambda integration out of the box
   • FastAPI service for distributed architectures
   • Event-driven architecture support

4. Enterprise-Grade Features

   • Type safety and validation
   • Comprehensive logging
   • Plugin system for extensibility
   • Workflow composition
   • Mathematical and data processing capabilities

Installation

Python

pip install agent-functions

JavaScript

npm install agent-functions

Quick Start

Python

from agent_functions import AgentFunction, workflow

@AgentFunction(
    category="nlp",
    description="Analyze text sentiment",
    agent_triggers=["sentiment_analysis_requested"]
)
def analyze_sentiment(text: str) -> dict:
    # Your sentiment analysis logic here
    return {"score": 0.8}

@workflow
def process_text(text: str):
    sentiment = analyze_sentiment(text)
    return {"result": sentiment}

JavaScript

import { AgentFunction, workflow } from 'agent-functions';

const analyzeSentiment = AgentFunction({
    category: 'nlp',
    description: 'Analyze text sentiment',
    agentTriggers: ['sentiment_analysis_requested']
})(
    async function(text) {
        // Your sentiment analysis logic here
        return { score: 0.8 };
    }
);

const processText = workflow(
    async function(text) {
        const sentiment = await analyzeSentiment(text);
        return { result: sentiment };
    }
);

Key Features

• Agent-First Design: Built from the ground up for AI agent integration
• Cross-Language Support: Python and JavaScript interoperability
• Cloud-Ready: Serverless and microservices support
• Type-Safe: Strong typing and validation
• Modular: Plugin system for easy extension
• Workflow Composition: Chain functions together
• Rich Analytics: Mathematical and data processing tools included

Use Cases

1. AI Agent Systems: Create function libraries that agents can discover and use
2. Microservices: Build event-driven services that work across languages
3. Data Processing: Compose complex data transformation workflows
4. Machine Learning: Create modular ML pipelines
5. API Development: Build flexible, agent-friendly APIs

Documentation

For full documentation, visit our documentation site

Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

License

MIT License - feel free to use this in your projects!