Two peas, one pod
Basic Memory vs Qdrant
By Sophie Filer · checked against both servers on 27 Oct 2026
Both give an assistant a memory that lasts between chats. Basic Memory stores it as ordinary Markdown files you can open, edit and keep in git, linked into a graph. Qdrant stores text as vectors and finds it by meaning, which suits a large pile of snippets better than a set of notes you curate.
Side by side
| Compared | Basic Memory | Qdrant |
|---|---|---|
| Made by | Basic Machines | Qdrant |
| Stored as | Markdown files in a folder | Vectors in a Qdrant collection |
| Finding things | Full-text search and links between notes | Similarity search by meaning |
| Tools | Many: write, edit, move, search, build context | Two: store and find |
| Edit by hand | Yes, in any editor or Obsidian | Not practical |
| Read-only switch | No; keep the folder in git | QDRANT_READ_ONLY=true |
| Tools | 27 (21 only read) | 2 (1 only read) |
| Context its tools take | At least ~3,200 tokens | – |
| Runs on | Your computer | Your computer |
| You need | uv (for uvx) | uv (for uvx); A Qdrant instance, or a local path for embedded storage |
| Our install check | Started in our test | Needs your setup first |
| Licence | AGPL-3.0-or-later | Apache-2.0 |
| GitHub stars | 4.4K | 1.6K |
| Last commit | 26 Oct 2026 | 4 Sept 2026 |
Stars and commit dates refresh daily; tool counts come from our own run of each server.
Pick Basic Memory if
- You want to read and correct what the assistant remembers
- You keep project notes or a personal knowledge base
- You want memory that stays as files on your machine
Pick Qdrant if
- You store many short snippets and look them up by meaning
- You want two simple tools rather than a note system
- You already run Qdrant, locally or in the cloud
Can you run both?
Two memory servers in one client tend to split what the assistant remembers. Choose one, or use Qdrant as a read-only snippet library next to Basic Memory for notes.