# Montycat

> Montycat is a self-hosted, Rust-powered NoSQL + vector database with built-in AI semantic search — the retrieval and memory layer for RAG, AI agents, and LLM apps. Records and their vector embeddings live in one engine, with on-device embeddings and no external vector database or embedding API. A self-hosted, open-source alternative to Pinecone, Weaviate, Chroma, and Qdrant, with official clients for Python, TypeScript/JavaScript, Dart/Flutter, and Rust.

## Capabilities

- [Meow Desktop](https://montygovernance.com/desktop): Released native visual database GUI for Montycat on macOS, Windows, and Linux. It manages connections, stores, keyspaces, records, schemas, governance, snapshots, and semantic search; detects a missing or outdated engine; and checks for Meow updates. Downloads are user-approved and open in the native operating-system installer.
- [Vector Database](https://montygovernance.com/vector-database): Self-hosted vector database with a NoSQL store in the same engine; kNN/HNSW search over on-device embeddings.
- [Semantic Search](https://montygovernance.com/semantic-search): Rank data by meaning with built-in vector search, by exact term with BM25 keyword search, or by both at once with hybrid ranking fused through reciprocal rank fusion. Indexed metadata, timestamp, pointer, and schema filters narrow the candidate set in every mode — no separate search service.
- [RAG Database](https://montygovernance.com/rag): Retrieval-augmented generation retrieval layer; embed-on-write, retrieve by meaning.
- [AI Memory](https://montygovernance.com/ai-memory): Long-term, semantically searchable memory for AI agents and LLMs that survives restarts.
- [Montycat MCP](https://montygovernance.com/mcp-server): Version 1.1.4 of the Model Context Protocol server exposes a Montycat engine as shared, persistent memory for AI agents. Its 24 tools cover memory, semantic/keyword/hybrid recall, live collaboration, governance, and read-only enforced-schema inspection for correctly typed structured writes and filtered retrieval. Claude, Codex, Cursor, and any MCP client can run it locally over stdio against a self-hosted engine.
- [Data Mesh](https://montygovernance.com/data-mesh): Each keyspace is an independently owned, domain-oriented data product.
- [Pinecone Alternative](https://montygovernance.com/pinecone-alternative): Self-hosted, open-source alternative to Pinecone, Weaviate, Chroma, and Qdrant.
- [Qdrant Alternative](https://montygovernance.com/qdrant-alternative): Montycat vs Qdrant — both Rust and self-hosted; Montycat combines vector search with a NoSQL store in one engine.

## Docs

- [Full documentation (single file)](https://montygovernance.com/llms-full.txt): The complete Montycat docs — installation, CLI, connection, keyspaces, schemas, CRUD, semantic, keyword, and hybrid search, subscriptions, snapshots, credentials, data-mesh governance, and advanced ops — as one plain-text document.
- [Markdown documentation index](https://montygovernance.com/docs/index.md): Individual Markdown documents for each topic, suitable for agents that only need one section.
- [Documentation](https://montygovernance.com/docs/installation): Install, connect, CRUD, schemas, subscriptions, semantic search, and data-mesh governance. Each section is a static page under /docs/<section>: installation, how-to-get-started, store-keyspace, schema, pointers, timestamps, insertion, retrieval, lookup, semantic-search, update, delete, expiration, subscriptions, snapshots, credentials, governance, advanced, security.
- [Governance — Markdown](https://montygovernance.com/docs/governance.md): Governance capabilities, scopes, qualifiers, grants, denials, inspection, auditing, and policy manifests.
- [Credentials — Markdown](https://montygovernance.com/docs/credentials.md): Superowner and owner credentials, data-access grants, and revocation.
- [Semantic search — Markdown](https://montygovernance.com/docs/semantic-search.md): Semantic configuration, queries, models, keyword and hybrid ranking modes, metadata filters, and precomputed vectors.
- [Download](https://montygovernance.com/download): Montycat Engine and Meow Desktop packages for Linux, macOS, and Windows.
- [Release notes](https://montygovernance.com/release-notes): Version history.

## Clients & distribution

- [Python client (PyPI)](https://pypi.org/project/montycat/)
- [Node.js / TypeScript client (npm)](https://www.npmjs.com/package/montycat)
- [Dart / Flutter client (pub.dev)](https://pub.dev/packages/montycat)
- [Rust client (crates.io)](https://crates.io/crates/montycat)
- [MCP server 1.1.4 (PyPI)](https://pypi.org/project/montycat-mcp/)
- [MCP server — Claude Desktop extension (.mcpb)](https://github.com/MontyGovernance/montycat-mcp/releases/latest/download/montycat-mcp.mcpb)
- [MCP server — generic latest release alias](https://github.com/MontyGovernance/montycat-mcp/releases/latest/download/latest.mcpb): The same current Claude Desktop extension published as `latest.mcpb` for automated consumers.
- [Engine (Docker Hub)](https://hub.docker.com/r/montygovernance/montycat)
- [Source (GitHub)](https://github.com/MontyGovernance)

## Getting started

Run the semantic edition (semantic search on by default):

```
docker run -d --name montycat \
  -p 21210:21210 -p 21211:21211 \
  -e MONTYCAT_SUPEROWNER="admin" \
  -e MONTYCAT_PASSWORD="change-me" \
  -v montycat_data:/var/lib/.montycat \
  montygovernance/montycat:semantic
```

On Apple Silicon use the `arm64-semantic` tag instead; `semantic` is the amd64 image.