memoripy vs memvid
AI agent memory projects compared
memoripy and memvid both provide persistent memory and retrieval for AI applications. memoripy emphasizes evidence screening, scoped records, and cited recall, while memvid packages content and search structures in a portable single file.
memoripy: Evidence-Tracked Memory for AI Agents
Memoripy is a local Python memory runtime for AI agents that screens incoming evidence, tracks changing facts over time, and explains recall with citations. It suits teams that need auditable, scoped memory rather than a simple vector store.
memvid: Store Persistent AI Memory in a Single File
Memvid is a Rust-based memory layer that stores content, search indexes, and metadata in a portable .mv2 file. It suits developers building agents that need local, persistent retrieval without operating a separate database service.
| memoripy | memvid | |
|---|---|---|
| Language | Python | Rust |
| License | Apache-2.0 | Apache-2.0 |
| Stars | 694 | 16.6k |
| Forks | 59 | 1.4k |
| Last analyzed | Oct 9, 2026 | Oct 3, 2026 |
Key differences
- memoripy is a Python memory runtime with admission policies, evidence tracking, temporal records, and retrieval citations; memvid is a Rust-based system centered on `.mv2` files and searchable content.
- memoripy offers independent retrieval lanes and scopes for users, agents, runs, projects, organizations, and namespaces; memvid provides full-text BM25 search and optional vector similarity search.
- memoripy includes CLI audits and APIs for correction, explanation, and versioned deletion; memvid uses append-only Smart Frames and supports inspection of earlier states.
- memoripy's core needs no external model or database, with optional PostgreSQL and service integrations; memvid is designed to avoid a separate database service, while embeddings may require model downloads or an API key.
- memoripy requires Python 3.10 or later and describes v4 as under development; memvid requires Rust 1.85.0 or newer and also lists Python, Node.js, and CLI interfaces.
- Both projects use the Apache-2.0 license; the supplied figures list 694 stars for memoripy and 16.6k for memvid.
Choose memoripy if you…
- need evidence screening, citations, and auditable changes to memory.
- want scoped retrieval and separate current and historical fact queries.
- plan to use a Python core without requiring a database or external model.
Choose memvid if you…
- want memory content and search structures packaged in a portable `.mv2` file.
- need full-text search with optional vector similarity search.
- want Rust, Python, or Node.js interfaces and optional media-processing capabilities.
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.