---
title: instibrain vs Basic Memory: which one fits your team
description: Both keep Markdown your AI reads over MCP. Basic Memory is open source and free to self-host; instibrain sets readers per space and seals what we cannot read.
url: https://instibrain.ai/compare/basic-memory
updated: 2026-11-10
site: https://instibrain.ai/llms.txt
---

# instibrain and Basic Memory, side by side.

Both keep Markdown your AI reads over MCP. Basic Memory gives you the code and the install. A brain gives a team control of who reads each page, the host included.

Basic Memory is the better choice for  The code, the install, and your documents in it.
- open source
- free on your own machine
- a knowledge graph
- document import
- a published team price
a brain adds  Who reads each page, the host included.
- readers per space
- sealed from the host
- a check on every save
- specialists a team shares
- rooms where AIs talk

first, what Basic Memory does better

## Basic Memory wins on five counts.

If one of these decides it for you, choose Basic Memory.

- Open source Under AGPL-3.0. Our source is not published.
- Free, on your machine No host at all. We host the brain.
- A knowledge graph It links your notes. A brain draws no graph.
- Document import PDF, Word, PowerPoint and CSV. A brain imports none.
- A published team price From $19 a month for two. A brain is free for one person; a team is by request.

the differences

## Seven differences, a picture each.

On the left, Basic Memory. On the right, a brain.

1

### Can you read the code, and run it yourself?

Basic Memory

- open source, AGPL-3.0
- free on your own machine
- your own servers, on Enterprise
Yes. The code is open, and it runs where you choose. Basic Memory does this better     a brain

- source not published
- hosted by us
- every page a file on your computers
No. We host it; the files stay with you.

**The detail**

Basic MemoryBasic Memory is open source under AGPL-3.0. It runs free on your own machine, in its hosted cloud, or on your own servers on its Enterprise plan.

a brainThe brain's source is not published, and you cannot run the brain on your own servers. The sync client keeps every page as a Markdown file on your computers, so the pages stay with you if we do not. Linked computers.
2

### Who reads what?

Basic Memory

one workspace · five roles
- owner
- admin
- editor
- viewer

Invite-only projects, with access per member, are on its Business plan.

Roles in a workspace. Private projects on Business.    a brain

team
- everyone
finance
- finance
legal · sealed
- two people
Each space has its own readers.

**The detail**

Basic MemoryA workspace has five roles: Owner, Admin, User Admin, Editor and Viewer. Invite-only projects, with access set per member, and audit logs are on its Business plan.

a brainEach space has named readers, editors and managers, and it is shared whole: no folder inside it is shared on its own. An AI reads only the spaces its person gives it. Spaces.
3

### What can the host read?

Basic Memory

```
in its cloud
encrypted in transit and at rest
on your own machine
nothing leaves it
```

Its cloud encrypts at rest. Self-hosted, there is no host.    a brain

```
sealed, on our server
rT9b8QDUyii5i+4ftgFBFd0C1brw==
UqEW2qJ84uwjJFQu4hKyH3oELpNmxG3n==
```

A sealed space reaches us as ciphertext.

**The detail**

Basic MemoryIts Security page says traffic is encrypted over TLS and files and their index are encrypted at rest; it names no end-to-end encryption. On your own machine, the notes never leave it, and no host reads anything.

a brainAn open space is not hidden from us: our server reads it, which is how a web AI and a phone read it. A sealed space is encrypted on the computers of the people it is shared with, and we hold ciphertext, so no web AI or phone reads it. We still see its size and timing. The limits, in full.
4

### Which AI reads it?

Basic Memory

- Claude
- ChatGPT
- Codex
- Cursor
- any MCP client
Claude, ChatGPT, Codex, Cursor, and more. It names ChatGPT and Cursor; we have not run them     a brain

- Claude Code
- Codex
- Claude on the web
- Claude on the phone
- ChatGPT, Cursor: not yet
The AIs that have run on our brain.

**The detail**

Basic MemoryBasic Memory's pages name Claude, ChatGPT, Codex and Cursor, and any MCP client.

a brainAny AI that speaks MCP connects to the brain's address. Run on our brain: Claude Code on Mac, Windows and Linux, Codex, and Claude on the web and the phone. ChatGPT and Cursor speak MCP and have not run on ours yet, so we promise no result for them. Connect your AI.
5

### How does it find a note?

Basic Memory

- semantic search
- a knowledge graph

The graph links one note to the next.

Semantic search, and a knowledge graph. A graph is something a brain does not have     a brain

- by words
- by meaning
- a knowledge graph
- inside a sealed space
By words and by meaning, over open pages.

**The detail**

Basic MemoryBasic Memory searches by meaning and builds a knowledge graph that links the notes.

a brainThe brain's search finds a page by its words and by its meaning, the best first, over open pages. It draws no knowledge graph, and it does not reach a sealed space, which our server cannot read.
6

### Does it take your documents?

Basic Memory

- PDF
- Word
- PowerPoint
- CSV

Each becomes a note your AI can search.

Yes. It imports them as notes. Basic Memory does this better     a brain

-  a page says where the file is
-  your AI opens it there

Through the store's own connector.

No. Pages are Markdown, and files stay put.

**The detail**

Basic MemoryBasic Memory imports PDF, Word, PowerPoint and CSV files, and makes a searchable note of each.

a braininstibrain imports no documents. Your files stay in your own systems; a page says which system holds what, and your AI opens the file through that store's own connector. Context, not data.
7

### What keeps it true?

Basic Memory

- a version on every save
- daily snapshots

In its cloud.

Versions and snapshots.    a brain

```
the check's answer to a save, an example
line 5  history: "we decided"
line 9  a long sentence, 34 words
line 12 a sentence pricing/floors.md holds already
```

A check on every save, which refuses nothing.

**The detail**

Basic MemoryIts cloud keeps a version on every save and takes daily snapshots.

a brainA check at every save flags a line that is too long, vague, written as history, or repeated from another page. It refuses nothing, so a person still decides. Every version names its writer. The check, on a real page.

side by side

## The whole comparison in one table.

Each row is a difference above; its detail is there.

| The question | Basic Memory | instibrain |
| Source | Open, under AGPL-3.0 | Not published |
| Where it runs | Your machine, its cloud, on-prem | Our server; files on your computers |
| Who reads what | Roles; private projects on Business | Readers per space |
| What the host reads | Encrypted at rest; none self-hosted | Open spaces; never a sealed one |
| Which AI reads it | Claude, ChatGPT, Codex, Cursor | Claude Code, Codex, Claude |
| Finding a note | Semantic search, a knowledge graph | Words and meaning; no graph |
| Documents | PDF, Word, PowerPoint, CSV | None; pages are Markdown |
| Keeping it true | Versions and snapshots | A check on every save |
| Running an AI | Hosted agents, by usage credits | None; you bring your AI |
| Price | Free; cloud from $19 a month | Free for one person; a team by request |

The facts about Basic Memory are its own, read on 10 November 2026: its home page, pricing, business and security pages. Check them there before you decide.

which one

## Choose by what you cannot give up.

Basic Memory gives you the code and the install. A brain gives a team control of its readers.

choose instibrain if A team has to trust what the pages say, and who reads them.
-  Readers, editors and managers are set for each space
-  Some knowledge must be closed to the host that stores it
-  Pages are kept true: a check at every save, and one page for one fact
-  Your team shares one method as a specialist, and its AIs talk in rooms

choose Basic Memory if You want the code, the install, and your documents in it.
-  You want open source you can read, change and run
-  You want a free install on your own machine, with no host at all
-  You want a knowledge graph across your notes
-  You want PDFs, Word files and slides imported as notes
-  You want a published price for a team today

where they part

## Three things a team gets from a brain, each with its cost.

None is free of a cost, so the cost is beside it.

-
Readers set per space The cost: no folder inside a space is shared on its own
-
A space the host cannot read The cost: no web AI reads it, no phone, and our search does not reach it
-
Pages kept true The cost: the check refuses nothing, so a person still decides

questions

## What people ask who know Basic Memory already.

**Is instibrain open source?**

No. The source is not published. If you have to read the code that holds your notes, Basic Memory is the one to choose.

**Can I run instibrain on my own servers?**

No. We host the brain. The sync client keeps every page as a Markdown file on your computers, so the pages stay with you if we do not.

**Can I move my Basic Memory notes into instibrain?**

They are Markdown files. Put the folder in a space on your computer, and it syncs as pages. instibrain builds no knowledge graph from them.

**What does a sealed space cost me?**

It opens only on the computers of the people it is shared with. No web AI reads it, no phone, and our search does not reach it. Our server still sees its size and timing. The limits, in full.

**What does instibrain not do?**

It runs no AI and imports no documents. We host the brain: you cannot run it on your own servers, and its source is not published. A personal brain is free, up to 200 pages and 20 MB; a team's price is by request.

## Give every AI your team uses the same pages.

Start with a free personal brain. For a team's account, write to us.

Sign up free  →
compare  instibrain vs Markdown in Git

A repo of your own, or readers per space and every AI.
compare  instibrain vs Obsidian

One person's vault, or a team's pages for every AI.
compare  All comparisons

What instibrain does not do, said first.

---

instibrain is made by WebWork AI, Inc. Everyone uses the same models · the difference is what yours knows.
