# ClaudioDrews/memory-os

A 7-layer memory operating system for Hermes Agent — persistent memory with Qdrant, structured facts, fabric recall, auto-curated wiki, and surgical context injection. Runs locally, any LLM provider.

Repository: https://github.com/ClaudioDrews/memory-os
Canonical: https://ross.abutalabs.com/products/memory-os
Language: Python
License: MIT
License Family: permissive
Topics: ai-memory, context-injection, docker, ground-truth, hermes-agent, local-first, open-source, persistent-memory, qdrant, rag, self-hosted, vector-database
Last push: 2026-06-10T10:40:43+00:00

## Health v2 (maintenance only)
Score: 52/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 86, release rhythm 35, longevity 6
- inputs: {"age_days": 94, "days_push": 84, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1338, forks 128 (observed 2026-08-28T04:04:25.736700+00:00)

## What it is
Memory OS is a seven-layer persistent memory system for the Hermes Agent, combining Qdrant vector search, structured facts with trust scoring, an auto-curated wiki, and surgical context injection. It runs entirely locally via Docker and works with any LLM provider, avoiding cloud lock-in.

## Use cases
- give my ai agent persistent memory across sessions
- stop repeating context to my coding agent every conversation
- self-hosted memory for llm agents with vector search
- inject relevant past decisions into agent context automatically
- store structured facts about my projects with trust scoring
- semantic search over all past agent conversations
- local-first agent memory without vendor lock-in

## When to choose
- you use Hermes Agent and want long-term memory that survives sessions
- you need local-first memory infrastructure with no cloud dependency or subscription
- you want provider-agnostic memory that works with OpenAI, Anthropic, OpenRouter, or Ollama
- you need token-efficient, surgical context injection rather than dumping full history

## When to avoid
- you don't use Hermes Agent or a compatible plugin system
- you want a fully managed cloud memory service with zero setup
- you can't run Docker services locally
- you need a turnkey memory solution for non-agent chatbots without integration work

## Facets
- artifact type: plugin
- maturity: active
- function: rag, vector-database, search-engine, agent-framework, mcp, chatbot
- domain: artificial-intelligence, large-language-models, self-hosted, developer-tools
- platform: self-hosted, python, cross-platform
- tags: persistent-memory, context-injection, qdrant, hermes-agent, local-first, memory-layers, llm-provider-agnostic, knowledge-wiki, ai-agents, retrieval-augmented-generation, docker, linux, macos

## Member repositories
- ClaudioDrews/memory-os (main) score 52

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:25.736700+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 2026-08-30T04:43:57.766384+00:00, confidence not recorded.
  - readme: https://github.com/ClaudioDrews/memory-os (fetched 2026-08-28T04:04:25.736700+00:00, sha 46ec59f3ff80)
- Data as of 2026-08-30T08:39:29.467469+00:00.