mozilla-ai/any-agent
A single interface to use and evaluate different agent frameworks
What it solves
It provides a unified interface for interacting with and evaluating multiple different AI agent frameworks. This eliminates the need to write framework-specific code when switching between or comparing different agentic systems like LangChain, LlamaIndex, or OpenAI Agents SDK.
How it works
Any-agent acts as a wrapper that exposes a common API. By using a standard AgentConfig and the AnyAgent.create method, users can instantiate agents from various supported frameworks (such as Agno, Google ADK, and smolagents) while maintaining the same code structure for running queries and managing tools.
Who it’s for
Developers and researchers who need to run, compare, or evaluate agents across multiple frameworks without being locked into a single ecosystem.
Highlights
- Unified API: Use a single set of imports and configuration patterns regardless of the underlying framework.
- Broad Framework Support: Compatible with TinyAgent, Google ADK, LangChain, LlamaIndex, OpenAI Agents SDK, smolagents, and Agno AI.
- Evaluation Tools: Includes capabilities to evaluate agent performance across different frameworks.
- Extensible: Supports integrating Model Context Protocol (MCP) tools and implementing multi-agent systems via an "Agents-As-Tools" approach.
Related
- Project
- Project
- Project
- Project
- Project