AgentIndex icon
AgentIndex
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Submit Tool
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Home/
Multi-Agent/
Sage
Sage logo

Sage

Active·★ 1.2k·MIT·Updated 2026-11-17

Sage is an RSI (Recursive Self-Improvement) agent harness that runs agents, grades them with an independent evaluator and tunes them into new candidate versions with recorded evidence. It ships as a desktop app, a multi-user server and the SAgents Python SDK.

This write-up is based on the project's README and public repository data. We have not hand-tested the tool.

#agents#ai#flutter#llm#manus#mcp#multi-agent#python
$ Install
$ git clone https://github.com/ZHangZHengEric/Sage.git && cd Sage && python -m pip install -e '.[desktop]'
↗ Visit site★ GitHub
01

Features

01RSI loop of runner, independent evaluator and tuner, keeping each version and its evidence
02Define instructions, Skills, tools and resources in `sage.yaml`
03Native tools, reusable Skills and MCP services with host-managed authorization
04Stateful sessions with streaming and pause/resume for human approval
05Multi-agent workflows with replaceable model, memory, storage and scheduling providers
02

Why choose it

+
+
+
+
03

Trade-offs

!
!
!
!
04

Maintenance check

ActiveChecked: 2026-11-12
Last commit
2026-11-12
Commits, last 90 days
100+
Releases, last 90 days
3
Latest release
v2.1.1 (2026-10-19)
Open issues
10
Recent issues with at least one comment
9/14
Commits per week, last 12 weeks
Commits per week, last 12 weeks: 31, 72, 31, 15, 37, 16, 48, 125, 95, 25, 21, 51

Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.

Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.

Most discussed open issues

  • #327 First RSI walkthrough: scripted example fails before JSON and documented result contract is stale2 comments · 2026-11-08
  • #325 v2 MCP: durable session pool and resilience for the MCP tool bridge1 comments · 2026-11-08
  • #317 Cross-link suggestion: Independent third-party evaluation1 comments · 2026-10-28
  • #315 Proposal: optional You.com web search provider in mcp_servers/search1 comments · 2026-10-23
  • #311 Add Cheaper Inference provider preset1 comments · 2026-10-21

Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.

05

Compatibility

macOS
Desktop app
Verified via docs
Windows
Desktop app
Verified via docs
Linux
Desktop app
Verified via docs
06

Quick start

1
$ git clone https://github.com/ZHangZHengEric/Sage.git
2
$ cd Sage
3
$ python -m pip install -e '.[desktop]'
07

Use cases

↳Improving an agent against a goal or curated usage history
↳Running configurable agents on the desktop with local models or providers
↳Embedding agents in a Python application through SAgents
08

Alternatives

⚡Top VS:
VSLightAgent★1.2k→VShive★11.2k→VSharness-sdk★9.2k→VSOpenExecutive★5.8k→
LightAgent logo
LightAgent★ 1.2k
LightAgent: Lightweight Python framework for OpenAI-compatible agents with tools, memory, guardrails, tracing, lifecycle hooks, multi-agent collaboration, and workflows.
VSCompare→
hive logo
hive★ 11.2k
Multi-Agent Harness for Production AI
VSCompare→
harness-sdk logo
harness-sdk★ 9.2k
Build an agent harness and control it end-to-end. Open-source SDK for production AI agents in Python & TypeScript - any model, any cloud.
VSCompare→
OpenExecutive logo
OpenExecutive★ 5.8k
AI-powered virtual executive team — a single coherent executive persona backed by 8 specialist agents (FastAPI + Next.js).
VSCompare→
DeepMCPAgent logo
DeepMCPAgent★ 897
Model-agnostic plug-n-play LangChain/LangGraph agents powered entirely by MCP tools over HTTP/SSE.
edsl logo
edsl★ 521
Design, conduct and analyze results of AI-powered surveys and experiments. Simulate social science and market research with large numbers of AI agents and LLMs.
ai-doc-gen logo
ai-doc-gen★ 771
AI-powered multi-agent system that automatically analyzes codebases and generates comprehensive documentation. Features GitLab integration, concurrent processing, and multiple LLM support for better code understanding and developer onboarding.
ask-user-questions-mcp logo
ask-user-questions-mcp★ 147
Better 'AskUserQuestion' - A lightweight MCP server/OpenCode plugin/Agent Skills + CLI tool that allows your LLMs ask questions to you. Be the human in the human-in-the-loop!
VSCompare→
See all alternatives →

Comments

Log in to leave a comment

No comments yet. Be the first!

On this page
01Features02Why choose it03Trade-offs04Maintenance check05Compatibility06Quick start07Use cases08Alternatives
ZHangZHengEric /
Sage
↗

Recursive Self-Improvement (RSI) agent harness evaluating, tuning, and evolving agents through evidence

1.2kMIT
Last commit1d ago
StatusActive
CategoryMulti-Agent
Links
Documentation↗Discussion↗Issues↗Releases↗
Cloud Infrastructure
Ad

Deploy on DigitalOcean

Host your MCP servers, Docker containers, and AI agents with $200 free credit over 60 days.

Claim $200 Free Credit
© 2026 AgentIndex.app|Built by a 10-year iOS Developer.
Best by use caseAboutContactAdvertisePrivacy PolicyTerms of UseQYSGitHubBuy me a coffee ☕

Browse by Category

MCP ServersCode AssistantWorkflow AutomationRAG / Knowledge BaseMulti-AgentBrowser AutomationLLM InfraDev ToolingObservability

Not affiliated with Anthropic, OpenAI or Microsoft.