Giving Claude Code persistent memory with a self-hosted MCP server
Claude memory without token costs, but requires running five services for one feature.
Your AI agents never start from zero again. Local-first MCP runtime with persistent memory across sessions and tools.
Another MCP memory server when several already exist.
Developers using AI coding assistants like Cursor or Copilot
Continue · Cursor · Other MCP memory servers
Claude memory without token costs, but requires running five services for one feature.
Claude Code memory without Redis—138 lines of diskcache beats 'use a vector DB' conventional wisdom.
Finally remembers your architecture decisions between Claude Code sessions — CLAUDE.md couldn't do this.
Claude Code memory server that auto-reads OAT tokens, routes LLM ops to local or cloud models.
Runs extraction and search server-side so your local MCP is a tiny HTTP client — no local DBs, no giant RAM leaks, and an easy npx install and .mcp.json or global MCP registration. It exposes clear tools (save_memory, recall_memories, extract_memories, get_project_context) and adds project-scoped + global preferences — a pragmatic fix for Claude Code's tiny flat-file memory. The tradeoff is obvious: usefulness depends on the hosted API (privacy, uptime, cost), and the repo looks early-stage with minimal commits and docs beyond the quickstart.
Hierarchical memory that persists across Claude Code, Cursor, and Windsurf—solve context amnesia.