The Project Brain

The project brain is the center of MindMux. It is not a notebook, not a transcript log, and not a vector database. It is the durable, structured memory of a project.

A concrete mental model

The simplest way to think about the brain is:

• root docs describe the whole project
• pages capture specific stable knowledge units

In practice, a project brain often looks like this:

brain/
├── background.md
├── architecture.md
├── flow.md
├── mindmap.md
├── stack.md
├── roadmap.md
├── index.md
├── schema.md
├── mindmux.json
├── pages/
│   ├── runtime-profile-model.md
│   └── task-execution-model.md

What lives in the brain

The brain has two primary layers:

• root docs
• pages

Root docs

Root docs are the fixed project-wide outputs:

• background
• architecture
• flow
• mindmap
• stack
• roadmap

These are the long-lived overview documents for the project. They are updated in place rather than multiplied into many variants.

Each one plays a different role:

• background: what the project is, for whom, and why it exists
• architecture: the main system structure
• flow: the important end-to-end user or system flows
• mindmap: the feature or capability map
• stack: major technical choices and open decisions
• roadmap: near-term and medium-term direction

Pages

Pages are the unbounded structured knowledge layer. Each page represents a distinct entity such as:

• a decision
• a concept
• a project thread
• a person
• a reference

Each page has two important parts:

• compiled_truth: the current best understanding
• timeline: the evidence and change history

This split is one of MindMux's most important ideas. The app keeps both the current conclusion and the path that led there.

That means a page answers two different questions:

• "What do we currently believe?"
• "How did we get here?"

Page lifecycle status

Every page also carries a lifecycle status:

• draft — still taking shape
• active — current project knowledge
• archived — no longer active, but kept for history

By default, ordinary reads and searches only include active pages. This keeps the working context focused. Draft and archived pages are still stored in brain/pages/ and can be surfaced explicitly when needed.

Promotion path

The intended knowledge flow is:

Not every piece of information needs to pass through every stage, but the structure shows the intended progression from raw material to durable project truth. A new page may start as draft while it is taking shape and move to active once it represents current project truth.

Why Markdown matters

MindMux uses Markdown and simple metadata because the brain is supposed to remain:

• portable
• inspectable
• versionable
• editable outside the app

That is a product choice, not just a storage implementation detail.

What the brain is not

The brain is not intended to be:

• a dumping ground for every transient idea
• a file upload bucket
• a replacement for source control
• an opaque AI memory store that only the product can read

How teams should think about it

The right mental model is:

• chat is where thinking happens
• the brain is where stable understanding lands

If a discussion would be expensive to reconstruct later, it probably belongs in the brain in some form.