ZHangZHengEric/Sage has 1,212 GitHub stars and 115 forks. GitHot measured 0 stars in 24 hours. Last checked 2026-11-09 07:39 UTC. Last push 2026-11-02.
An RSI Agent Harness for running, evaluating, and evolving agents—with tools, skills, MCP, stateful execution, and multi-agent workflows.
What you can do
- 📦 Packages and versions - Define instructions, Skills, tools, and resources in sage.yaml; retain candidates, diffs, and evaluation records.
- 🧠 Models and context - Connect model providers and configure routes, memory, and context budgets.
- 🔌 Tools, Skills, and MCP - Compose native tools, reusable Skills, and MCP services with host-managed authorization.
- 💾 Stateful execution - Persist Sessions, stream output, and pause/resume for human input and approvals.
Measured star growth
- in 24h
- 0
- in 7 days
- Update delayed
- in 15 days
- Update delayed
- in 1 month
- Update delayed
Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.
docs/en/README.md
Originallayout: default title: Overview nav_order: 1 lang: en ref: v2-README permalink: /en/
{% include lang_switcher.html %}
Overview
Sage is an RSI Agent Harness, bringing task execution, independent evaluation, and recursive self-improvement into one runtime. These docs explain supported usage, integrations, and operating boundaries.
Choose your entry point
| Your goal | Start here |
|---|---|
| Understand how RSI improves a Package | RSI workflow |
| Inspect an iteration without a model key | First RSI run |
| Install Desktop and use RSI | Desktop |
| Embed Sage in a Python application | Quick start · Host development |
| Provide a multi-user web service | Server |
| Configure models, tools, and Skills | Configuration · Tools and MCP |
| Integrate an existing client | Python API · HTTP API |
Understand and extend the Harness
Core concepts · Architecture · Plugins · Memory · Support boundaries.
Operate and contribute
Environment variables · Troubleshooting · Development · Model-provider acceptance.
Running from source requires Python 3.12+. Desktop installers bundle the backend; Server currently supports one worker.
Documentation scope
This site maintains Sage product, SDK, protocol, and operating references. Legacy usage is covered by Sage v1 and the legacy reference directory. Customer delivery proposals, temporary audits, and phase-specific validation notes are not retained as repository documentation.
Explore related projects
affaan-m/ECC
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
300,313 starsNousResearch/hermes-agent
The agent that grows with you
262,134 starsmultica-ai/andrej-karpathy-skills
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
231,460 starsfirecrawl/firecrawl
Supercharge your AI agents with data from the web and beyond. Building the library for superintelligence. 🔥
200,178 stars