https://github.com/lupantech/AgentFlow
AgentFlow: In-the-Flow Agentic System Optimization for Effective Planning and Tool Use
https://github.com/lupantech/AgentFlow
agentic-ai agentic-systems llms llms-reasoning multi-agent-systems reinforcement-learning tool-augmented
Last synced: about 2 months ago
JSON representation
AgentFlow: In-the-Flow Agentic System Optimization for Effective Planning and Tool Use
- Host: GitHub
- URL: https://github.com/lupantech/AgentFlow
- Owner: lupantech
- License: mit
- Created: 2025-09-27T02:17:05.000Z (3 months ago)
- Default Branch: main
- Last Pushed: 2025-10-08T03:23:49.000Z (2 months ago)
- Last Synced: 2025-10-08T03:27:47.753Z (2 months ago)
- Topics: agentic-ai, agentic-systems, llms, llms-reasoning, multi-agent-systems, reinforcement-learning, tool-augmented
- Language: Python
- Homepage: https://agentflow.stanford.edu
- Size: 5.7 MB
- Stars: 2
- Watchers: 1
- Forks: 1
- Open Issues: 0
Awesome Lists containing this project
- awesome-llm-agents - AgentFlow - Trainable multi-agent framework (Frameworks)
- Awesome-AI-Agents - AgentFlow - Trainable multi-agent framework coordinating planner, executor, verifier, generator via in-the-flow optimization.  (Frameworks / Advanced Components)
- awesome_ai_agents - AgentFlow - A trainable multi-agent framework that coordinates four specialized modules (planner, executor, verifier, generator) through in-the-flow optimization, achieving significant performance gains over monolithic approaches by directly training the planner within multi-turn task loops using Flow-GRPO reinforcement learning. Any tools can be smoothly integrated into this framework, math, coding, scientific, search, finance, news, etc. [github](https://github.com/lupantech/AgentFlow) | [website](https://agentflow.stanford.edu/) | [paper](https://huggingface.co/papers/2510.05592) (Building / Frameworks)