Product

Mind Archive — Your Personal AI Mind Archive. Own your AI memory. Simple, private, and yours.

The problem

People are accumulating a large and growing amount of genuinely valuable thinking inside AI chat products: research, decisions, drafts, explanations, plans, code, personal reflection.

That material has three problems.

1. It lives on someone else’s server. You can usually export it, but you cannot really use it where it is.
2. It is trapped per provider. Move from ChatGPT to Claude and you leave your history behind.
3. It is not readable on its own terms. A provider export is a large JSON file, not something a person can browse, search or keep.

What Mind Archive does

Mind Archive gives you a personal archive of your AI conversations and related knowledge that lives on your own computer, in formats you can read, and that survives any change of provider.

It lets you:

• Import exports from AI providers, starting with ChatGPT
• Preserve them as human-readable Markdown and JSON on your filesystem
• Organise them with projects, tags and metadata
• Search across everything
• Export the whole archive at any time
• Optionally synchronise to storage you choose — never by default

Who it is for

People who use AI assistants seriously and want to keep what comes out of them: researchers, writers, engineers, students, consultants, and anyone who has ever thought “I know I worked this out with an AI six months ago.”

The first version targets desktop and laptop use. Mobile is not a Milestone 1 target, though the interface is built responsively where practical.

Principles

Local first. The application works fully with no network, no account and no third-party service. This is not a degraded offline mode; it is the normal mode.

Privacy first. Nothing leaves your machine unless you explicitly configure it. Cloud features are off by default and cannot be enabled silently.

You own your data. Your content is stored as ordinary files you can read, grep, back up and move. The database indexes your archive; it never owns it. If Mind Archive disappeared tomorrow, your archive would still be readable.

AI-provider agnostic. No provider sits at the architectural centre. ChatGPT is the first importer target and nothing more.

Human readable. The interface, the documentation and the stored data are written for people. “Import archive”, not “Initialize ingestion pipeline”.

Simple. A single primary workspace with minimal navigation. No dashboards, no permanent sidebars, no unnecessary animation. It should feel like a calm, modern document archive.

Proprietary, and honest about it. Mind Archive is All Rights Reserved: not open source, not source-available, no licence granted. What that does not change is the promise underneath — your archive is ordinary readable files on your own disk, exportable at any time, whatever happens to the software. See LICENSING.md.

What Mind Archive is not

• Not a chat interface. It archives conversations; it does not host them.
• Not a cloud service. There is no hosted product and no account.
• Not an AI provider client. It does not need an API key to do its core job.
• Not a note-taking app competitor. It is an archive first.

Interface direction

Two appearance choices sit in the header, and both are remembered. A palette — Light minimal (the default), Warm paper, or Ink & violet — and a theme: Light, Dark, or System, which follows the computer. All six combinations must look good. See DECISIONS.md D-029.

Each is a menu button rather than a row of options, which is what lets the header be a single row: brand, navigation, then actions and appearance. Below 1024px the navigation folds into a menu; the appearance buttons never do, and lose only their text label. Nothing on either surface scrolls sideways — wide content scrolls inside its own container instead. See D-037.

The interface should read as something a thoughtful person designed: calm, clean, obvious, lightweight, professional. It should specifically not look AI-generated — no gradient-heavy card grids, no decorative dashboards, no jargon.

Settings must be simple, and only settings that genuinely work may exist. No placeholder toggles.

Milestones

Milestone 1 delivers a runnable foundation. Milestone 2 delivers the ChatGPT importer — the first point at which Mind Archive becomes useful. Later milestones add browsing and search, organisation, more importers, and finally optional cloud synchronisation.

The full plan is in ROADMAP.md, with the authoritative scope in MILESTONES.md.