# P0luz/Ombre-Brain

Repository: https://github.com/P0luz/Ombre-Brain
Canonical: https://ross.abutalabs.com/products/ombre-brain
Language: Python
License: MIT
License Family: permissive
Last push: 2026-08-25T15:46:03+00:00

## Health v2 (maintenance only)
Score: 79/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 93, longevity 10
- inputs: {"age_days": 140, "days_push": 8, "days_rel": 51, "gap_med": 1, "n_releases_24m": 6}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1262, forks 872 (observed 2026-08-28T04:04:10.143147+00:00)

## What it is
Ombre Brain is a long-term emotional memory system for Claude and other MCP clients, exposing 16 MCP tools over a single /mcp connector. Memories are tagged with Russell valence/arousal coordinates, stored as Obsidian-compatible Markdown with YAML frontmatter, retrieved via hybrid keyword (rapidfuzz + BM25) and vector semantic search, and decayed by an improved Ebbinghaus forgetting curve.

## Use cases
- give Claude persistent memory across conversations
- store chat memories with emotional valence and arousal coordinates
- search past memories with hybrid keyword and vector semantic retrieval
- let old memories naturally decay and archive with a forgetting curve
- import ChatGPT, Claude, or DeepSeek chat history into a memory store
- browse and manage memory buckets in Obsidian or a web dashboard
- expose a remote MCP memory server with OAuth 2.1 over HTTPS

## When to choose
- you want an MCP-native persistent memory layer for Claude or another MCP client
- you prefer human-readable Markdown/Obsidian storage over a plain vector database
- you need emotionally weighted, decaying memories rather than static key-value recall
- you want resilient writes where embeddings are generated in a background queue without blocking persistence

## When to avoid
- you need hard deletion or GDPR-style physical erasure of memories, since deletion only archives files
- you want a simple key-value or pure vector store without emotional/decay semantics
- you need a memory system for non-MCP clients without REST/API integration work
- you require a battle-tested enterprise memory solution rather than a niche community project

## Facets
- artifact type: service
- maturity: active
- function: vector-database, search-engine, mcp, rag, database, sdk
- domain: artificial-intelligence, large-language-models, chatbots, databases, self-hosted, developer-tools
- platform: python, self-hosted, cross-platform
- tags: mcp-server, long-term-memory, emotional-memory, obsidian, forgetting-curve, hybrid-retrieval, claude, valence-arousal, memory-management, dashboard, ai-agents, retrieval-augmented-generation, docker, web-server

## Member repositories
- P0luz/Ombre-Brain (main) score 79

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:10.143147+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 2026-08-30T05:04:12.864639+00:00, confidence not recorded.
  - readme: https://github.com/P0luz/Ombre-Brain (fetched 2026-08-28T04:04:10.143147+00:00, sha 23812dbccd7f)
- Data as of 2026-08-30T08:39:29.467469+00:00.