musubi
Active·★ 11·Apache-2.0·Updated 2026-10-06
Musubi is a specialized memory server designed for fleets of AI agents, enabling them to share learned information and collaborate effectively. It features a unique three-plane memory model and a lifecycle engine that matures raw observations into a human-reviewable knowledge base.
#AI Agent Memory#Knowledge Base#Lifecycle Management#Vector Database#LLM Inference#Multi-Agent System#Python Service#Obsidian Integration
01
Features
01Three-Plane Memory Model (Episodic, Concept, Curated)
02Knowledge Lifecycle Management with Maturation and Curation
03Shared Agent-to-Agent Communication ("Thoughts" Plane)
04Hybrid Vector Search and Reranking (Qdrant, BGE-M3, SPLADE)
05Obsidian Vault Integration for Human Review
02
Compatibility
Python
Runtime
Verified via docs
Docker
Deployment
Verified via docs
Qdrant
Vector Database
Verified via docs
Ollama
LLM Runtime
Verified via docs
Obsidian
Knowledge Base
Verified via docs
03
Quick start
1
$ git clone https://github.com/ericmey/musubi
2
$ cd musubi
3
$ docker compose -f quickstart/docker-compose.yml up -d --wait
04
Use cases
↳Collaborative AI Agent Fleets requiring shared context and memory.
↳Automated knowledge base generation and maintenance, maturing raw data into structured insights.
↳Human-in-the-loop knowledge refinement through Obsidian integration.
↳Advanced information retrieval for agents, supporting semantic search beyond keywords.
↳Enabling specialized agents (e.g., note-taking, Q&A, data cleanup) to leverage collective intelligence.
05
Alternatives
ruflo★ 78.4k
🌊 The leading agent orchestration platform for Claude. Deploy intelligent multi-agent swarms, coordinate autonomous workflows, and build conversational AI systems. Features enterprise-grade architecture, distributed swarm intelligence, RAG integration, and native Claude Code / Codex Integration
neo-agent-brain★ 11
Point a cross-model AI engineering team at your own codebases. The Agent OS keeps named AI maintainers' identity, memory and review across labs alive between sessions — the Brain Neo.mjs runs on itself. Multi-tenant: onboarding a codebase is a config entry, not a fork.
engram-memory★ 38
The highest-scoring AI memory system ever benchmarked that isn't reliant on LLM reranking. And it's free & burns less tokens.
lavern★ 315
An agentic law firm. Yours. 67 specialist AI agents that review documents through evidence-backed debate, with mandatory human gates and a 10-pass verification loop. Apache 2.0.
forge-orchestrator★ 162
NXTG-Forge orchestrator: multi-AI task orchestration with file locking, knowledge capture and drift detection. Rust. Public preview (source-available, FSL-1.1-ALv2).
claude-vibe-squad★ 164
Multi-model AI orchestration where behaviour is Markdown, not code. One coordinator routes scoped task packets to 71 role-based specialists across 5 model families (Codex / Claude / Gemini / Grok / Kimi) — each working in an isolated git worktree, each reviewed by a rival family before it lands. One tmux session, no server.
Comments
Log in to leave a comment
No comments yet. Be the first!