gentle-ai
Gentle-AI configures the AI coding agents you already use: Claude Code, Cursor, OpenCode, Codex, Pi, and more. Choose persistent memory, Organic-Driven Developm
What is gentle-ai?
Gentle-AI configures the AI coding agents you already use: Claude Code, Cursor, OpenCode, Codex, Pi, and more. Choose persistent memory, Organic-Driven Development, curated skills, MCP servers, personas, and optional bounded review. Open source, no agent lock-in.
Overview of gentle-ai
What it is
Gentle-AI is an open-source tool that configures AI coding agents such as Claude Code, Cursor, OpenCode, Codex, Pi, and others. It provides persistent memory, a workflow (Organic-Driven Development), curated skills, MCP servers, personas, and optional bounded review. It aims to give agents memory, a workflow, and evidence of their work.
Key features
- Engram: persistent project context that accumulates across sessions.
- Organic Driven Development (ODD): keeps small changes lightweight and larger work recoverable with feature documents.
- Delegate for a reason: subagents are used only when specific criteria (exploration, parallel writers, context backstop, independent verifier) are met.
- Measured routing: benchmarks show proportional routing reduced time and cost for small tasks.
- Test-first by default: capture failing behavior test before implementation.
- Receipt-Driven Development (RDD): optional review mode that freezes the candidate before review.
- Deterministic by design: the gentle-ai binary owns review transitions and ODD guidance.
- Works with 17 agents: Pi, OpenCode, Claude Code, Codex, Cursor, VS Code Copilot, Gemini CLI, Kilo Code, Kimi Code, Kiro IDE, Qwen Code, Hermes, Antigravity, Windsurf, OpenClaw, Trae, Conductor.
- Additional components: skills library, Context7 MCP, CodeGraph, security deny-list, config backups, doctor, personas, themes, model assignment.
Use cases
- Developers who want AI agents to remember project context across sessions.
- Teams that need a consistent workflow for AI-assisted coding with evidence of changes.
- Projects requiring test-first development and review processes.
- Users of multiple AI coding agents who want a shared configuration and workflow.
Getting started
bash # macOS (Homebrew) brew install gentleman-programming/tap/gentle-ai # macOS / Linux (curl) curl -fsSL https://raw.githubusercontent.com/Gentleman-Programming/gentle-ai/main/scripts/install.sh | bash # Windows (PowerShell) — source install …
License and pricing
Licensed under MIT. Open source.
Based on the project's public README and repository metadata.