# Basic Memory

> Basic Memory is an MCP server that lets Claude, Codex, Cursor, ChatGPT and other MCP clients read and write a knowledge graph of plain Markdown notes (observations and wikilinks) on your disk, indexed in SQLite with hybrid full-text and FastEmbed vector search. It needs no LLM of its own, but you need an MCP-capable AI client to use it. Code is AGPL-3.0. Installed with uv (Python). A paid hosted version ($15/month per the README) adds sync and mobile.

- Repository: https://github.com/basicmachines-co/basic-memory
- Official docs: https://github.com/basicmachines-co/basic-memory#readme
- Checked: 2026-10-16

## What it replaces
- Mem0 Platform: Memory for coding agents, Knowledge graph memory (partial). Memory is Markdown notes the AI writes through MCP tools, not automatic extraction from every conversation.

## Facts, with sources

| Question | Status | Answer | Source | Checked |
|---|---|---|---|---|
| Runs on my machine? | Confirmed | Yes, local files plus a SQLite index; air-gapped use supported | https://github.com/basicmachines-co/basic-memory#why-basic-memory | 2026-10-15 |
| Code licence | Partial | AGPL-3.0 (copyleft) | https://github.com/basicmachines-co/basic-memory/blob/main/LICENSE | 2026-10-15 |
| Model weights licence | Confirmed | No separate model weights (optional FastEmbed models download on first use) | https://github.com/basicmachines-co/basic-memory#optional-cross-encoder-reranking | 2026-10-15 |
| Commercial use | Partial | Yes, under AGPL-3.0 copyleft terms (share source of modified versions offered over a network) | https://github.com/basicmachines-co/basic-memory/blob/main/LICENSE | 2026-10-15 |
| Paid services needed? | Confirmed | None required by Basic Memory; you need an MCP-capable AI client, which may itself be paid | https://github.com/basicmachines-co/basic-memory#cloud-vs-local | 2026-10-15 |
| Hardware | Unknown | Not documented | https://github.com/basicmachines-co/basic-memory#readme | 2026-10-15 |
| Operating systems | Partial | Windows x86\_64, Linux x86\_64/ARM64 and macOS are listed in the semantic-search compatibility table; Intel Macs need an ONNX Runtime pin or OpenAI embeddings for local semantic search | https://docs.basicmemory.com/concepts/semantic-search#platform-compatibility | 2026-10-16 |
| Setup effort | Confirmed | Easy, uv tool install plus adding the MCP server to your AI client | https://github.com/basicmachines-co/basic-memory#get-started | 2026-10-15 |

GitHub: 4240 stars, 321 forks, last push 2026-10-16, archived: no.
---
Source: RepoQuarry (https://repoquarry.dev/projects/basic-memory). Data licence: CC BY 4.0. Quotes are from project documentation and belong to their authors.