Basic Memory provides a user‑owned knowledge base that stores information as plain Markdown files enriched with a searchable knowledge‑graph index. The platform combines full‑text, semantic, and structured search, allowing AI tools to query notes by meaning, metadata, or typed relationships while keeping the underlying files editable and portable.
Funding
Funding not disclosed
Founders
Product
Problem
AI assistants often rely on proprietary, opaque memory stores that lock user data into black‑box vector databases, making the information unreadable, non‑portable, and vulnerable to loss when switching providers. This prevents users from editing, versioning, or reusing their knowledge across different AI tools.
Solution
Basic Memory offers a user‑owned knowledge base built on plain‑text Markdown files enriched with a full‑text, semantic, and knowledge‑graph index. The system keeps the underlying files editable and portable while providing instant, meaning‑based search and structured queries via front‑matter filters. Typed relations and observations create a traversable graph that AI agents can query to retrieve contextually relevant information. Integrations with major LLMs (Claude, ChatGPT, Gemini, Codex) and developer tools (Obsidian, VS Code, Cursor) enable a single source of truth that works across platforms. The product is available as a free, open‑source local version and as a paid cloud service with private sync, team workspaces, and enterprise controls.
Target Audience
Primary users are developers, knowledge workers, and teams who need a transparent, editable knowledge base that can be accessed by multiple AI assistants and collaboration tools.
Features
- Plain Markdown files with built‑in front‑matter support, allowing any text editor or version‑control system to manage the data
- Real‑time semantic search using vector and hybrid indexing to retrieve notes by meaning rather than keywords
- Structured search and metadata filtering on custom front‑matter fields for spreadsheet‑like queries
- Typed relations and observations forming a knowledge graph that AI agents can traverse for context‑aware responses
- Drop‑in agent integrations (Claude, ChatGPT, Gemini, Codex, etc.) via the Model Context Protocol for search‑before‑answer recall
- Multi‑project and shared team workspaces with per‑project graphs and real‑time collaboration
- Open‑source AGPL‑3.0 codebase with 3K+ stars, enabling self‑hosting, auditability, and community contributions