Zero-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
A memory layer for AI agents — the one that already runs an autonomous research OS over ~6,000 notes.
Memory is the mother of the Muses. An agent with no memory has no ideas.
pip install agora-mnemo · PyPI · Hugging Face · DOI 10.5281/zenodo.21128549 · Homepage · MIT · v1.7.0
mnemo is the recall + consolidation core of Agora — an autonomous research system — distilled into a single file with no required dependencies. It does the four things agent memory actually needs, the way that held up running in production for weeks.
Most "agent memory" libraries are demos. This one is extracted from a system that has used it daily to curate a 6,000-note knowledge base, and whose consolidation behaviour we have measured, not assumed (see Provenance below).
That is the whole loop most agents need: remember, recall, correct, and audit — in one zero-dependency file. Add embed=yourmodel for semantic recall; everything below is depth (governance, poison-resistance, bitemporal, multi-tenancy) you can reach for when you need it.
Runnable examples live in examples/: basics · correction & erasure · semantic recall.
Jump to: Correction (measured) · Governance & erasure · Install · MCP server · The four operations · Five rules · Provenance & receipts · Threat model
Correction is a first-class operation (measured across systems)
Any memory layer can store a fact and retrieve it. The harder, less-benchmarked property is integrity: when a fact is corrected, can the store undo the correction on command, and does restating a retired value resurrect it? mnemo treats correction as a first-class
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/DanceNitra--mnemo.html)
Self-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
AlexanderBoger/SaveContextGit LFS for your LLM context — 20x Context Savings — an MCP server that keeps big documents out of the model and returns byte-exact quotes for a fraction of the tokens.
mienetic/mnema🧠 Mnema — Long-term memory for AI via MCP × Vector DB. Pluggable backends, hybrid search, Auto Dream consolidation, CLI + REST API + SDK.
hung12ct/culiSelf-improving context & memory for Claude Code and OpenAI Codex. One canonical knowledge store injected only when relevant — hooks push budgeted context, an MCP server serves depth on demand, and it
bybit-exchange/kaasTurn scattered notes, docs and transcripts into a queryable Markdown wiki — an LLM knowledge-base compiler with MCP access, no embeddings, self-hosted.
ozankasikci/global-agent-memoryLocal-first, project-aware memory MCP server for Claude Code, Codex, and other AI agents, with Obsidian and an owner dashboard.
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.