aipype 0.2.0a1

A modular AI agent framework with declarative pipeline-based task orchestration

aipype

Core AI agent framework with declarative pipeline orchestration.

Installation

pip install aipype

Quick Start

Use the @task decorator for clean, Pythonic agent definitions:

from aipype import PipelineAgent, task, llm, search

class ResearchAgent(PipelineAgent):

    @task
    def find_sources(self) -> dict:
        """Search for articles - no dependencies, runs first."""
        return search(self.config["topic"], max_results=5)

    @task
    def summarize(self, find_sources: dict) -> str:
        """Summarize sources - dependency auto-inferred from parameter."""
        return llm(
            prompt=f"Summarize: {find_sources}",
            model="gpt-4o",
            temperature=0.3
        )

agent = ResearchAgent("research", {"topic": "AI trends"})
agent.run()
agent.display_results()

Key Concepts:

• @task decorator marks methods as pipeline tasks
• Parameter names matching task names create automatic dependencies
• self.config provides access to agent configuration
• Helper functions: llm(), search(), mcp_server(), transform()

Environment Variables

# LLM Providers
export OPENAI_API_KEY=sk-your-key-here
export GEMINI_API_KEY=your-key-here
export OLLAMA_API_BASE=http://localhost:11434

# Search
export SERPER_API_KEY=your-serper-key-here

Development

Requirements

• Python ≥3.12