memxt
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Install
curl -fsSL https://raw.githubusercontent.com/Yupcha/memxt/main/install.sh | bash~/.memxt/bin/memxt demo # 60-second tour on a throwaway palace (the GIF above)
~/.memxt/bin/memxt adopt --write # wire your agents + mine this repo
~/.memxt/bin/memxt doctor # if anything's off — every ✗ prints its fix commandmacOS / Linux, x86_64 & arm64. One ~7 MB binary + an on-device MiniLM model. No accounts, no keys, zero network at query time.
Related MCP server: CozoDB Memory MCP Server
Claude Code
/plugin marketplace add Yupcha/memxt
/plugin install memxtDone. Every session now wakes up with project decisions and recent work (~10 ms), auto-saves before compaction and on stop, and recalls via memory_search → memory_get. /remember and /recall when you want to be explicit.
Other agents — one shared palace
Agent | Setup |
Codex |
|
Cursor |
|
Grok CLI |
|
Any MCP client |
|
All of them read and write the same MEMXT_DB. Open a new session after wiring. Full guide → docs/harnesses.md.
How it works
Wings (projects) → Rooms (topics) → Drawers (verbatim memory + embedding)
│
search = vectors + FTS5 + facts + recency, fused
wake-up = identity + project truth + recent work, ~10 msMemories are kept verbatim (no LLM rewriting), searched hybrid (semantic + exact keywords like 0x5C), and consolidated by dream: hot vectors stay f32, cold history compresses to 4-bit, stale facts expire.
Numbers (Apple Silicon, method in BENCHMARK.md): wake-up ~10 ms · warm search sub-ms · ~100 MB peak RAM · continuity bench 6/6 · 72% fewer tokens than re-pasting docs each session (method).
Built for agents (v0.4)
Procedural memory — repeated successful workflows become procedures;
memxt skills --emitturns them into Claude Code skills.Grounded memory — memories anchor to file + content hash; recall tags
[stale]when the evidence changed.Token-budget recall —
budget_tokenson search/wake-up returns the best brief that fits.Fleet-ready — parallel subagents share one palace: attributed writes, lock-retry, scratch tier with
memory_promote.Usage-learned relevance — memories the agent actually opens rank higher; ignored ones decay.
Sleep-time compute —
memxt dream --daemonconsolidates in the background and flags contradicting facts.Client-model sampling — opt-in (
MEMXT_SAMPLING=1): fact extraction by your client's own model, still $0.
vs the alternatives
We studied 15 agent-memory products (full teardown). The field splits into cloud memory APIs (Mem0, Supermemory), Python/Node local layers (MemPalace, agentmemory, MemOS), and Claude-first plugins (claude-mem). memxt's bet: the only native-binary, $0-memory-loop option that's deep in every coding harness.
memxt | claude-mem | Mem0 / Supermemory | |
Code leaves machine | never | compresses via cloud LLM | yes, by default |
Memory LLM bill | $0 | metered | metered |
Works with | Claude · Codex · Cursor · Grok | Claude-first | platform |
Remembers | verbatim + facts + anchors | AI summaries | extracted entities |
Nobody else has: grounded memories that admit they're [stale], procedural memory that emits agent skills, token-budget recall, or a 60-second demo you can verify on your own repo.
CLI
memxt demo [--keep] 60-second tour on a throwaway palace
memxt doctor Self-diagnosis with exact fix commands
memxt adopt [--write] [--no-mine] Wire up agents + optionally mine the repo
memxt mine <path> [wing] Incremental codebase ingest
memxt search <q> --mode hybrid | memories | documents | facts | episodes
memxt wake-up [--budget N] | inspect | dream | serve
memxt dream --daemon | --status | --contradictions
memxt skills [--emit [dir]] Procedural memory → Claude Code skills
memxt anchors [--verify] Grounded-memory anchor health
memxt forget | export | import | mcp | hook | instructionsMEMXT_DB=~/.memxt/palace.db MEMXT_MODEL=~/.memxt/lib/minilm.gguf
MEMXT_WING=my-project # optional; defaults to the git-root nameBuild from source
Zig 0.16 + cmake:
git clone --recursive https://github.com/Yupcha/memxt && cd memxt && zig build --release=fastROADMAP.md · MIT · 3D logo
If memxt saves you one re-explain session, star it.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Persistent memory for Claude Code and Cursor. Stop re-explaining your project every session.
Private memory for AI agents: save and recall facts across Claude Code, Codex, Cursor and ChatGPT.
Persistent memory for AI agents — shared .brain across Claude Code, Cursor & Codex
One memory for your AI agents and your team: Claude, ChatGPT, Cursor and your agents share it.
Related MCP Servers
- AlicenseAqualityDmaintenancePersistent memory for Claude Code. Automatically indexes every conversation and provides production-grade hybrid search (BM25 + vectors + reranker) via MCP tools. 100% local, zero config, zero API keys, zero invoice.1650 npm8MIT
- AlicenseNot gradedqualityDmaintenanceLocal-first memory for Claude & AI agents with hybrid search, Graph-RAG, and time-travel, runs entirely on your machine.43 npm1Apache 2.0
- AlicenseBqualityBmaintenancePersistent, local memory for AI coding agents that learns how you work, not just what you said. Supports Claude Code, Codex CLI, Cursor, and any MCP client.7998 PyPI77MIT
- AlicenseNot gradedqualityBmaintenanceLocal-first memory for your AI agent. One SQLite file you own — offline, no API key. Plugs straight into Claude Code.Apache 2.0