# EverMind-AI/Raven

The Harness of Harnesses: a trusted, persistent, self-evolving multi-agent ecosystem for all-domain collaboration.

Repository: https://github.com/EverMind-AI/Raven
Canonical: https://ross.abutalabs.com/products/raven
Homepage: https://raven.evermind.ai
Language: Python
License: Apache-2.0
License Family: permissive
Topics: ai, ai-agents, anthropic, chatgpt, claude, codex, evermind, hermes, hermes-agent, llm, openai, openclaw, openhuman, self-evolving, self-improving
Last push: 2026-08-26T13:53:39+00:00

## Health v2 (maintenance only)
Score: 81/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 99, longevity 7
- inputs: {"age_days": 104, "days_push": 7, "days_rel": 8, "gap_med": 3, "n_releases_24m": 14}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3640, forks 69 (observed 2026-08-28T04:08:12.769822+00:00)

## What it is
Raven is an open-source, self-improving agent harness built on EverOS that combines terminal-first execution, local tracing, long-term memory, skills, evaluation, and reusable workflows for long-running AI agent work. It continuously improves through an evaluation-execution-verification-memory-feedback loop and composes specialized agent harnesses into an all-domain collaboration network.

## Use cases
- run long-running AI agent tasks with persistent memory
- build a self-improving personal AI agent that learns from past runs
- orchestrate multiple agents across different domains and models
- evaluate and benchmark agent performance on task quality and cost
- migrate existing agents and reuse a large library of built-in skills
- build custom agents for specific scenarios and share them

## When to choose
- you need a persistent, memory-first agent that improves over time
- you want terminal-first agent execution with local tracing and evaluation
- you need multi-model, multi-domain agent orchestration with verified capability routing
- you want an open-source alternative to closed agent harnesses with a large skill ecosystem

## When to avoid
- you need a simple one-shot chatbot without memory or self-evolution
- you require a fully mature, battle-tested production system with long-term support
- you want a lightweight agent library to embed in your own codebase rather than a full harness platform
- you are uncomfortable with a system that rewrites its own skills, runtime, and policies

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, llm-inference, rag, machine-learning, workflow-automation, cli
- domain: artificial-intelligence, large-language-models, developer-tools
- platform: python, windows, cli, cross-platform
- tags: agent-harness, self-improving, long-term-memory, multi-agent, everos, skills, evaluation, terminal-first, self-evolving, ai-agents, automation, macos, linux

## Member repositories
- EverMind-AI/Raven (main) score 81

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:12.769822+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:31:47.644681+00:00, confidence not recorded.
  - readme: https://github.com/EverMind-AI/Raven (fetched 2026-08-28T04:08:12.769822+00:00, sha 988df332551c)
  - homepage: https://raven.evermind.ai (fetched 2026-08-29T09:26:12.778472+00:00, sha 811341c3e460)
  - site_page: https://evermind.ai/faqs (fetched 2026-08-29T09:26:12.788346+00:00, sha 1649a2ffb662)
- Data as of 2026-08-30T08:39:29.467469+00:00.