v10.45 — Now Available

MCP Memory Service

Persistent memory for AI agents. Semantic search, knowledge graph, automatic consolidation — now accessible from Claude.ai via Streamable HTTP.

GitHub PyPI Package

What's New in v10.45

OpenAI-compatible quality scoring: point memory scoring at Ollama, LiteLLM, MLX-LM, or vLLM. Soft-delete UPDATE guards. ~1,758 tests.

Homelab Quality Scoring

Set MCP_QUALITY_AI_PROVIDER=openai-compatible to score memories with any OpenAI /v1/chat/completions-compatible endpoint — Ollama, LiteLLM, MLX-LM server, or vLLM. No cloud API key required. (PR #790)

Graceful Fallback Chain

New Tier 2 in the quality scoring chain: local ONNX → openai-compatible → Groq → Gemini → implicit signals. Endpoint failures fall through silently — no exception reaches the storage path.

📊

Soft-Delete UPDATE Guards

Seven UPDATE memories SET ... statements in sqlite_vec.py now include AND deleted_at IS NULL, preventing accidental modification of tombstoned rows. (PR #783, @filhocf)

How It Works

Claude.ai connects to your memory server through Streamable HTTP with OAuth 2.1 authentication.

Browser
Claude.ai
Transport
Streamable HTTP
Auth
OAuth 2.1 + PKCE
Server
MCP Memory
Storage
Knowledge Graph
Data flow Auth boundary

Built for Production

Battle-tested with comprehensive testing and optimized for performance.

0
Tests
0
AI Apps Supported
0
Read Latency
0
Cache Speedup

Get Started in Seconds

Install from PyPI and connect your AI agents to persistent memory.

Click to copy pip install mcp-memory-service