memvid vs RuVector
Rust-based persistent memory and retrieval projects compared
memvid and RuVector both provide Rust-based tools for persistent AI memory and local retrieval. memvid centers on a portable, searchable single file, while RuVector is a broader toolkit with separate packages for vector search, graph memory, learning, and agent integrations.

memvid: Give AI Agents Portable, Single-File Memory
Memvid is a Rust-based, single-file memory layer for AI agents, designed to replace complex RAG pipelines and server-based vector databases. It packages data and search indexes together, using Tantivy for full-text search and HNSW for vector search.

RuVector: Build Agent Memory with Rust and ONNX
RuVector is a Rust-based vector search and persistent agent-memory toolkit, with ONNX embeddings for local semantic retrieval. It offers Rust, Node.js, and browser/WASM paths, plus separate graph, learning, and MCP components.
| memvid | RuVector | |
|---|---|---|
| Language | Rust | Rust |
| License | Apache-2.0 | MIT |
| Stars | 17.3k | 4.6k |
| Forks | 1.5k | 613 |
| Last analyzed | Oct 30, 2026 | Oct 19, 2026 |
Key differences
- memvid packages content, metadata, and search structures in `.mv2` files; RuVector provides persistent vector storage and similarity search through its components.
- memvid combines BM25 full-text search with optional vector search; RuVector emphasizes vector search, metadata filtering, and HNSW or flat indexes.
- memvid lists Apache-2.0 licensing, while RuVector lists MIT licensing.
- memvid offers Rust, Python, Node.js, and CLI interfaces; RuVector also has browser/WASM, PostgreSQL, and MCP package surfaces.
- RuVector is a monorepo whose graph, learning, and MCP components require explicit integration; memvid focuses on memory files and their retrieval capabilities.
- memvid requires Rust 1.85 or newer for its Rust library; RuVector requires Rust 1.77 or newer for its core workspace, and Node.js 18 or newer for its npm path.
Choose memvid if you…
- need a portable, searchable memory file without a separate database service.
- want full-text BM25 retrieval alongside optional vector search in one format.
- need Python, Node.js, or Rust interfaces to a local memory layer.
Choose RuVector if you…
- want vector search with metadata filtering and HNSW or flat indexes.
- need to explore separate browser/WASM, PostgreSQL, or MCP package paths.
- plan to combine vector retrieval with graph or feedback-driven memory components.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.