Memo
Persistent, local-first semantic memory for AI coding agents.
About Memo
Memo is a local MCP server that gives AI coding agents like Claude Code, Cursor, and Windsurf persistent memory across sessions. It stores memories semantically using embeddings and vector search in a local SQLite database, allowing agents to recall architecture decisions, bug patterns, and project context without losing information between conversations. Every memory is automatically mirrored to an Obsidian-compatible vault, letting you browse and graph your memory store using Obsidian's full UI. The tool uses pure Go embeddings (GoMLX with BAAI/bge-small-en-v1.5 model) and sqlite-vec for cosine distance search, keeping everything on your machine with zero cloud dependencies. One limitation is that body edits made directly in Obsidian are silently overwritten on the next sync—you must use memo update to modify memories.
At a glance
- Platforms
- Web (via CLI/MCP), macOS, Linux, Windows, API (MCP server)
- API
- Available
- Integrations
- Claude Code, Cursor, Windsurf, Obsidian, Claude CLI (for LLM polish feature)
- Last verified
- June 2026
★Among the most viewed AI tools in Developer Tools — ranked #21 of 573.
♥Bookmarked by 2 people on Aigregator.
Who It's For
- •AI coding agents (Claude Code, Cursor, Windsurf, Codex)
- •Software engineers using AI-assisted coding workflows
- •Teams building AI agent memory systems
- •Developers wanting persistent context across projects
How It Works
- 1Runs as a local MCP server over stdio JSON-RPC, spawned as a child process by your AI agent
- 2Accepts memories via memo_remember tool with automatic deduplication (exact SHA256 hash or semantic similarity >= 0.90)
- 3Embeds memory content to 384-dimensional vectors using the BAAI/bge-small-en-v1.5 model (pure Go, no external dependencies)
- 4Stores vectors and metadata in SQLite with sqlite-vec extension for cosine KNN search
- 5Syncs every memory to a Markdown file in ~/.memo/vault with YAML frontmatter (Obsidian-compatible)
- 6Optionally polishes vault Markdown asynchronously using Claude CLI for richer formatting (callouts, wikilinks, headings)
- 7Provides nine MCP tools (remember, search, recall, list, similar, update, forget, reconcile, status) with millisecond response times
How to Use Memo
- 1Install via Homebrew: brew install ybonda/tap/memo or use the install script
- 2Connect to Claude Code with: claude mcp add --scope user memo -- memo serve
- 3For Cursor/Windsurf, add memo to .cursor/mcp.json with command: memo serve
- 4Use CLI commands directly: memo remember --content "..." --tags "tag1,tag2" to store memories
- 5Query with: memo search --query "topic" for semantic search or memo recall for formatted context
- 6Open ~/.memo/vault in Obsidian to browse and graph memories visually
- 7Wrap all memo MCP calls in a subagent (recommended in CLAUDE.md) to prevent context bloat
Key Features
- •Semantic memory search using vector embeddings (BAAI/bge-small-en-v1.5)
- •Automatic deduplication (exact hash + cosine similarity >= 0.90)
- •Configurable memory types (note, incident, ticket, guides, architecture, custom)
- •MCP server interface (memo serve) and CLI tools (memo remember, search, recall, etc.)
- •Obsidian-compatible vault with YAML frontmatter and automatic sync
- •LLM-polished markdown export using Claude CLI (optional, async)
- •Two-tier deduplication (exact SHA256 + semantic similarity)
- •Custom memory types validated at runtime
- •Full-text search and graph visualization via Obsidian integration
Use Cases
- •AI agents remembering API rate limits, rate limiting patterns, and security rules across projects
- •Storing and retrieving production incident post-mortems and bug patterns for faster debugging
- •Maintaining architecture decisions and system design patterns that persist across conversations
- •Building a project-agnostic knowledge base accessible to Claude Code, Cursor, or other MCP clients
- •Using Obsidian to visualize and explore memories with graph view and full-text search
Pros & Cons
Advantages
- •Pure Go implementation with zero external dependencies (no Python, ONNX, separate vector DB)
- •All data stays local in ~/.memo/memories.db—full privacy and offline capability
- •Obsidian integration provides rich UI (graph view, full-text search, mobile sync) at no extra cost
- •Seamless MCP integration means agents can automatically manage memories without explicit prompting
Disadvantages
- •Direct Obsidian edits to memory content are silently overwritten on next sync (only CLI/MCP updates persist)
- •Embedding model (~50MB) requires local download on first run; offline-only after that
- •LLM polish timeout (60s default) can fail on large memories; reverts to deterministic formatting
Alternatives
See all Memo alternatives →Related Tools
View All Tools
AI Code Converter
Convert code between languages with AI.
SQLAI.ai
AI-powered SQL query generator and database interaction tool.
AgentGPT
Assemble, configure, and deploy autonomous AI agents in your browser.
Amazon Q Developer
AI assistant for developers to build, test, and deploy applications faster.
AskCodi
AI code assistant for developers. Write, debug, and understand code faster.
AutoGPT
Autonomous AI agent leveraging GPT-4 for complex task management.
Reviews for Memo
Based on 2 reviews
Rating Distribution
tehila rabi
Mar 23, 2026
I use it with Cursor and sometimes with Claude Code (still learning it). I save anything that has importance and then in any session on any project I can fetch that context and it's super useful!
ybonda
Mar 15, 2026
I use this tool every day. What I like in it is the simplicity. The architecture is complex. but the interaction with the tool is super easy. in the end, all your your memories are stored with local sqlite file. so you can get them from any agent anytime via MCP
Frequently Asked Questions
What is Memo?
Memo is a local MCP server that gives AI coding agents like Claude Code, Cursor, and Windsurf persistent memory across sessions. It stores memories semantically using embeddings and vector search in a local SQLite database, allowing agents to recall architecture decisions, bug patterns, and project context without losing information between conversations.
How much does Memo cost?
Memo is free to use.
Is Memo free?
Yes, Memo offers a free plan you can start with.
What are the best Memo alternatives?
Popular Memo alternatives include Cursor AI, Bloop AI, Fig AI.
What is Memo used for?
Memo is commonly used for AI agents remembering API rate limits, rate limiting patterns, and security rules across projects, Storing and retrieving production incident post-mortems and bug patterns for faster debugging, Maintaining architecture decisions and system design patterns that persist across conversations.
Does Memo have an API?
Yes, Memo offers an API for developers.
What platforms does Memo support?
Memo is available on Web (via CLI/MCP), macOS, Linux, Windows, API (MCP server).
Information Accuracy
Please note: While we regularly update all tool information including descriptions, features, pricing, and other details, this information may change over time as tools evolve and update their offerings. For the most current and accurate information, we recommend visiting the official website directly. Our goal is to provide you with comprehensive and up-to-date information to help you make informed decisions.