AmritaCore

AmritaCore is a lightweight Agent runtime built on top of AmritaSense. It delivers native async streaming, tool integration, event hooks, and memory management — everything you need to build interactive, production‑ready Agent applications without the overhead of heavyweight frameworks.

🚀 Fast Lookup

import asyncio
from amrita_core import create_agent

async def main():
    agent = create_agent(
        base_url="https://api.openai.com/v1",
        api_key="your-api-key",
        model="gpt-4o-mini",
    )
    chat = agent.get_chatobject("Hello, how are you?")
    print(await chat.full_response())

asyncio.run(main())

🔑 Key Features

• Interactive‑first design — native async streaming with suspend/resume
• Vendor‑agnostic adapter system — OpenAI, Anthropic, and extensible
• Declarative dependency injection — type‑safe, based on function signatures
• Event‑driven hooks — intercept and modify the processing pipeline
• Tool system — @simple_tool, @on_tools, and MCP client support
• Advanced memory management — automatic context window and token optimisation

📖 Documentation

Full guides, API references, and examples at core.amritabot.com.

🤝 Contributing

We welcome contributions! Please see our contribution guidelines for more information.

📄 License

This project is licensed under the Apache 2.0 License — see the LICENSE file for details. All versions of AmritaCore are released under Apache 2.0.

Other files

• CONTRIBUTING.md — Contribution guidelines
• CODE_OF_CONDUCT.md — Code of conduct
• ZH-CN.md — 简体中文
• EN-US.md — English (US)

Unstable Features

• Python 3.14+ Supporting: we are not sure if it will work well on Python 3.14+(No GIL Version).