蒸留蔵 — distilled long-term memory for agents: recall by meaning, writing gated by evidence, one kura per agent mode. Ships as a DeepSeek Harness plugin and an MCP server.
A long-term memory for agents that is distilled, not accumulated. Recall works by meaning, writing is gated by evidence, and one server can hold several separate memories — one per agent mode — so switching mode switches what the agent remembers.
Ships as a DeepSeek Harness plugin, an MCP server for any other host, an HTTP service, and a Python library. Standard library only; no vector database, no embeddings, no framework.
Two failures kill an agent's long-term memory, and they kill it from opposite sides.
Retrieval by keyword misses the thing you needed. A question about "SSD inference chips" shares no word with a memory titled "running the 2.6T model off an SSD tier" — yet they are the same subject. Word search returns nothing; the agent answers from nowhere. The fix here is not embeddings but recognition: the entire index (one line per memory, written as a recognition trigger) goes into one prompt, and a small model names what bears on the question. An index of ~500 memories is around 6k tokens — a few percent of a modern context window, and it sits in the prefix cache.
Writing everything poisons the store. An agent asserts something; a naive distiller records the assertion as a fact; the next agent reads it back as ground truth and repeats it with more confidence. That loop is self-reinforcing, and prompt instructions do not stop it — measured, not assumed. So the write path is gated by deterministic Python: every candidate memory must carry quotes that exist character-for-character in the raw material, tagged with where they came from.
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/lna-lab--distill-kura.html)
Local-first long-term memory for AI agents - plain Markdown vault + thin MCP server. CJK-aware BM25, token-efficient by design.
DanceNitra/mnemoZero-dependency agent memory + MCP server. Value-ranked recall, consolidation, and a first-class correction & erasure channel (revert, lineage-aware retraction, tamper-evident receipts). Measured inte
Pseudogiant-xr/Pseudolife-MCPPersistent long-term memory for Claude — an MCP server with an associative memory bank, a canonical-fact cortex, sleep-like dream consolidation, and a web console. Not quite alive.
mienetic/mnema🧠 Mnema — Long-term memory for AI via MCP × Vector DB. Pluggable backends, hybrid search, Auto Dream consolidation, CLI + REST API + SDK.
17861102832/wanyimem为了服务于大众
rwnalds/engramSelf-hosted MCP server + dashboard giving Claude Code, Cursor & Hermes agents shared long-term memory over a git-backed markdown vault. Agent-native, Obsidian-compatible second brain — markdown RAG, n
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.