whereismyai
PocketFlow
PocketFlow is an open source minimalist LLM framework that implements core abstractions in just 100 lines.
Pocket Flow is an open source minimalist LLM framework consisting of just 100 lines of code. It captures the core abstraction of LLM frameworks using a graph structure, allowing developers to implement design patterns like multi-agents, workflows, and RAG. The project provides multiple versions in languages such as Typescript, Java, C++, Go, Rust, and PHP. It supports agentic coding where AI agents help build agents. The framework includes various tutorials and examples ranging from basic chat bots and structured output extraction to advanced coding agents.
PocketFlow at a glance
Key features
- Zero dependencies and zero vendor lock-in
- Minimalist 100-line core abstraction graph
- Support for multi-agents and workflows
- Retrieval-augmented generation implementation
- Available in Typescript, Java, C++, Go, Rust, and PHP
- Agentic coding support for AI development
Use it for
- Build conversational chat bots with history
- Extract structured data from resumes
- Generate code and test cases iteratively
- Process invoices and documents using vision
- Create automated research and writing agents
Alternatives to PocketFlow
- MLflow — Complete observability and evaluation for production AI apps.
- SmallCode — Runs fully local with budget-managed context and a forgiving parser for small models.
- CodeGraph — It eliminates file crawling, reducing AI agent tool calls by 88% and cutting costs.
- OpenCode — Works with 75+ model providers, local models, and existing subscriptions
- mini-SWE-agent — It offers high performance with a simplified, minimal codebase.
- Pi Pocket — Runs sessions that survive restarts and supports multiplayer device sharing.
- Flowise — Modular building blocks and visual tools to build and scale agentic systems fast.
- FlutterFlow — It allows for full code export to avoid vendor lock-in.
- LLM Space — Inspect every model call and tool run inside the agent loop locally.
- Stagewise — Maintain full sovereignty over AI models while running agents in parallel.
- OpenAPPA — Tracks data flows algebraically to prevent prompt injection without breaking agents.
- jevgrep — Reduces coding agent costs by providing targeted, relevant repository context.
All PocketFlow alternatives