The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Install
git clone https://github.com/Open-Curiosity/gini-agentInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →Gini Agent
Gini Agent is a personal agent that remembers, improves, and runs without forcing you to read a log line.
Gini is not just a chat box, CLI, messaging bot, or pile of tools. Chat is an interaction surface. The runtime is the system of record for conversations, runs, tasks, approvals, memory, skills, jobs, tools, traces, audit events, and runtime health.
It's also built to be a product, not plumbing. When the agent needs you, it hands you a purpose-built control right where the work is — a secure field, a choice, a sign-in handoff, a confirm-before-send — instead of a prose instruction or a detour. And because every client speaks the same runtime, those controls reach you on your phone as readily as the desktop, so the machine the agent runs on is not where you have to be.
Docs
- Whitepaper: the gaps this project is closing and the bar it's measured against
- Architecture Overview: gateway/client map
- Gateway And Control Plane: runtime process, BFF, auth, instances, ports, disk layout
- Conversation And Runs: chat, runs, tasks, plan steps, traces, and audit handoff
- Memory: retain, recall, embeddings, reranking, review, and storage
- Skill Learning From Skills: how Gini improves its own skills from task outcomes (two-tier reward, attribution, the daily review, the human gate)
- Runtime Capabilities: current CLI/API capability map and verification commands
- Model Providers: per-provider setup guides (credentials, prerequisites, CLI/web config) for OpenAI, Anthropic, Bedrock, Azure, OpenRouter, DeepSeek, Codex, and Local
- Operations: install, start, stop, smoke, diagnostics, and cleanup
- Docker Deployment: run an instance headless in a container with a real browser under Xvfb
- Remote Access: tunnel modes and confirmation, plus a self-contained guide per tunnel provider — Gini Relay, Tailscale, ngrok, Cloudflare — the same pages the app opens inline
- Releases: versioning, CHANGELOG conventions, and the release process
- Migrating from openclaw: import an existing openclaw install into gini
- Implementation Notes: source layout and module boundary rules
- Roadmap: shipped surfaces and what's planned, with design intent
Architecture decisions
- Architecture Decision Records: index of all ADRs and how to add new ones
Architecture In One Sentence
Gini's runtime is the gateway: a single Bun process per instance owns state and performs work. The Next.js web app, CLI, Expo mobile app, MCP surfaces, and messaging bridges are clients of the same authenticated /api/* contract. See Architecture Overview for the design.
Similar plugs
Reverse engineer anything with agents, from app behavior down to native binaries.
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype…
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io
Open-source Python library for AI browser agents, plus Browser Use Cloud: stealth browsers and a hosted agent API.