HomeProjectsArticlesPK BoardAbout
Explore Projects

PocketFlow

Normal
GitHub Python MIT

Description

PocketFlow is a minimalist 100-line LLM framework that lets Agents build Agents, enabling complex AI agent workflows through a clean abstraction layer.

Key Features

  • Only 100 lines of code, zero dependencies, zero vendor lock-in — ultra lightweight
  • Graph-based abstraction supporting Agent, Multi-Agent, Workflow, and RAG design patterns
  • Supports Agentic Coding — let AI agents like Cursor build agents for 10x productivity
  • Available in Python, TypeScript, Java, C++, Go, Rust, and PHP
  • Rich cookbook examples: Chat, Structured Output, Streaming, Guardrail, and more
  • Install via pip or copy the 100-line source directly — minimal onboarding friction

Use Cases

💡 Rapidly build chatbots and conversational systems with multi-turn history
💡 Create RAG (Retrieval-Augmented Generation) pipelines for knowledge-base Q&A
💡 Build multi-agent collaboration systems with inter-agent communication and workflow orchestration
💡 Use for education and prototyping — the simplest implementation of LLM framework core abstractions
💡 Extract structured data from unstructured text, such as parsing resume information

Strengths & Limitations

✅ Strengths

  • • High community interest (11.1k stars)
  • • Permissive open-source license (MIT)
  • • Responsive to issues, low backlog

Tags

agent-framework llm workflow minimalist python

Categories

🤖 Agent Framework

Quick Start

Install via pip install pocketflow, or copy the 100-line pocketflow/__init__.py directly from GitHub. Build Chat, RAG, and Agent apps quickly using cookbook examples. Docs at the-pocket.github.io/PocketFlow.
Visit GitHub Visit Website

Related Projects

AutoAgent

9.8k · Python
Stale B

Fully-automated and zero-code LLM agent framework that enables users to build and deploy custom AI agents through natural language without writing code.

agent-frameworkzero-codeautomation +2
  • · Fully-automated agent building through natural language — no coding required to create, customize, and deploy AI agents
  • · Zero-code framework democratizing AI development for users without technical backgrounds
  • · Self-managing workflow generation that dynamically creates and optimizes agent pipelines from high-level descriptions
View full details → View on GitHub

ART

10.7k · Python
Active A+

ART (Agent Reinforcement Trainer) trains multi-step agents for real-world tasks using GRPO reinforcement learning, enabling on-the-job training for models like Qwen, Llama, and more.

reinforcement-learningagent-trainingGRPO +3
  • · GRPO-based reinforcement learning for training multi-step agents on real-world tasks
  • · W&B Training serverless RL with 40% lower cost and 28% faster training via shared inference cluster
  • · Pre-built notebooks for email search, 2048, Tic Tac Toe, Codenames, and MCP tool mastery
View full details → View on GitHub

Hive

11.0k · Python
Active B

Hive is a production-ready multi-agent execution harness providing state management, failure recovery, observability, and human-in-the-loop control with auto-generated multi-agent topologies for complex business workflows.

multi-agentagent-frameworkproduction +3
  • · Zero-setup, model-agnostic execution harness that auto-generates multi-agent topologies for complex workflows
  • · Graph-based execution DAG for safe parallel task coordination with deterministic fault tolerance
  • · Persistent role-based memory that evolves with project context — no manual state management
View full details → View on GitHub