
Basic Memory
Basic Memory gives AI assistants durable, local-first memory by reading and writing structured Markdown notes, enabling reusable context across conversations and tools.

Basic Memory is a local-first “memory layer” that lets AI assistants build and reuse long-term context across chats. It stores knowledge as human-editable Markdown files and exposes that knowledge to compatible LLM clients via the Model Context Protocol (MCP).
Key Features
- Bi-directional read/write memory: AI can create and update notes, and you can edit them with standard tools
- Local Markdown storage with semantic patterns (frontmatter, observations, relations) to form a traversable knowledge graph
- Local indexing and search backed by SQLite for fast retrieval
- MCP server integration to connect with compatible AI clients (for example desktop assistants and editors)
- Multi-project organization for separate knowledge bases
- Optional sync workflows, including real-time syncing and cloud-oriented commands
Use Cases
- Build a personal knowledge base that persists across AI conversations without repeated re-explaining
- Maintain project “working memory” for coding, research, or writing using Markdown and wiki-style linking
- Share consistent prompts, instructions, and structured notes across different AI tools while keeping content editable
Limitations and Considerations
- Effectiveness depends on maintaining consistent note structure (observations/relations) for higher-quality retrieval
- Some cross-device features may depend on optional syncing workflows rather than the core local-only setup
Basic Memory is a practical way to turn conversations into durable, structured notes that both humans and AI can navigate. By keeping the source of truth in plain text Markdown, it aims to stay interoperable with existing editors and workflows while enabling richer, reusable AI context.