P
11.2k
Stars
+934
Stars/month
1
Commits (90d)
0
Releases (6m)
Star Growth
+2.8k (33.3%)estimated from velocity
Overview
PocketFlow is a lightweight, dependency-free framework for constructing AI agents, multi-agent systems, workflows, and RAG implementations. It is designed to be expressive while avoiding bloat and vendor lock-in, and promotes 'agentic coding' where AI assistants can build agents.
Deep Analysis
Key Differentiator
A fully functional agent framework distilled into approximately 100 lines of core Python code.
⚡ Capabilities
- • Agent creation
- • Multi-agent systems
- • Workflow orchestration
- • RAG implementation
🔗 Integrations
LLM providers (implied)Cursor AI (for agentic coding)
✓ Best For
- ✓ Developers seeking a minimalist agent framework
- ✓ Rapid prototyping of agent systems
- ✓ Educational understanding of agent architecture
✗ Not Ideal For
- ✗ Enterprise platforms needing extensive built-in tooling
- ✗ No-code agent building
⚠ Known Limitations
- ⚠ Minimalist by design, lacks extensive pre-built modules
- ⚠ Requires developer knowledge to implement
Alternatives
L
LLM Agents
Build agents which are controlled by LLMs
s
smolagents
🤗 smolagents: a barebones library for agents that think in code.
L
Langroid
Harness LLMs with Multi-Agent Programming
L
Lagent
A lightweight framework for building LLM-based agents
Compare PocketFlow
Maintain PocketFlow?
Show your live rank in your README, or put PocketFlow in front of every visitor to AgentoolRank.