manifest.md How it works FAQ About
manifest.md

Unified AI memory that you own.

Portable across tools. Private by design. Totally free.

Nothing to migrate — your context flows in as you use your tools.

"Of the 6.4 hours spent 'bot-sitting' per week, 2.3 … was feeding the AI context."
Work AI Institute, via the AI Daily Brief
Stop bot-sitting.
1 Build your context

Bring in what AI already knows about you

Pull it in from the tools you already use, or write it down. None of it is required to start — a few lines about who you are is a real manifest, and you can add the rest whenever. Everything on this page happens in your browser; nothing is sent anywhere until you choose where to keep it.

Bring your files

Drop your CLAUDE.md, AGENTS.md, .cursorrules — or any context file or notes you've built yourself.

And/or paste your context

Ask your AI tools to give you all the context they have on you, or copy the memory they keep in their settings (ChatGPT, Claude, and the rest) — then paste it here. Or use this space to add anything else you want every model to understand about you and what you need from them.

2 Build your Context OS

Choose how to keep it

Three ways, depending on you. Each one says plainly where your context lives.

Recommended

GitHub Connect

One click builds a private repo in your GitHub, with your context arranged into a structure Manifest designs for it — we call it a Context OS. In plain terms: it gives every AI tool you use the same context, tells each one to load only what it needs for what you're doing right now, and keeps itself current as you work, handing you updates to approve. More below.

What GitHub will ask: permission to write to just that one repo — and the moment setup is done, you can uninstall the app. Your Context OS is plain files in your GitHub; it keeps working without us.
No GitHub account? It's free — make one, then come back and click Connect. Your context is saved here.
Own each step

GitHub DIY

Hold everything, yourself. Download your full Context OS as a folder and put it in a private repo you create — same living context as Connect, built entirely in your browser. Nothing of yours ever passes through us.

Then: create a private GitHub repo, upload the folder, and use the Step 3 prompt.
3 Bring it everywhere you use AI

Bring it into any tool

Tell your tool to use the manifest.

GitHub

Paste this into any tool that can reach your repo — a coding tool opened in it, or a chat tool connected to your GitHub. Your tool reads your repo directly; nothing routes through us.

First time on a new repo, your tool sets up your context with you — a couple of questions, then your approval. After that, this just loads it. (Some tools need a nudge through the steps.)

4 End of session

Keep it current

When you wrap up, have the model save anything durable, so your context grows with you.

GitHub

Your Context OS already knows how to update itself; this just kicks it off, pointed at your exact repo.

It proposes the changes for your approval.
You hold your context. Nothing is ever sent to us.

Free, all of it. If it saves you time: $5 suggested donation.

How it works

Two things, plainly. Your manifest is your portable context — who you are, what you're working on, how you like to work, where you're headed. The name comes from the cargo manifest a ship carries port to port: one list of everything aboard. Yours travels tool to tool.

Your Context OS is that manifest plus the layer that makes it work everywhere — load levels so a model pulls only what a task needs, adapters so it behaves the same in Claude Code, Cursor, or ChatGPT, a privacy model, and an update loop that keeps it current. All plain markdown, human-readable, entirely yours.

Manifest builds it. You keep it.

Bring your context in

You don't start from a blank page — you start from what your tools already know. And none of it is required: a few sentences typed straight into the open field are a working manifest. Import the rest whenever you like, or never.

Pull your memory out of the chat tools you use. ChatGPT and Perplexity expose it in settings; with Claude or Gemini, ask "summarize everything you know about me and how I work" and copy the answer. Drop in any context files you've already written — CLAUDE.md, AGENTS.md, .cursorrules. Add anything else in one open field: notes, your résumé, whatever you'd want every model to know.

However it arrives, it lands exactly as you wrote it. Manifest captures your context; it never rewrites it.

Three ways to keep it

Each one says plainly where your context lives.

Connected (recommended). Authorize Manifest once and it writes the full Context OS into a private repo in your own GitHub — folders and all, in a click. Living instead of static: it loads by task, behaves per-tool, and updates as you work. The permission GitHub asks for covers just that one repo, and once your files are written you can uninstall the app on the spot — your Context OS is plain markdown in your GitHub and keeps working without us.

The link. One link that Claude and ChatGPT read directly — every new chat starts already knowing you, and the app can save new facts back as you go. No GitHub, no sign-in, no account. Your manifest is kept for you, encrypted, with the link itself as the only key; export or delete it anytime. How the link works.

Do it yourself. Download the same Context OS as a folder — built entirely in your browser — and put it in a private repo you create. You hold everything; nothing of yours ever passes through us.

How it organizes itself

Your imported files land in the repo whole, in a sources/ folder — no keyword parser guessing where things go.

The structure gets built the first time you point an AI tool at the repo. Your Context OS ships a SETUP.md that tells the tool to read everything in sources/ and sort it into the real files: who you are, what you're working on, your tech stack, a folder per project, private overlays for anything sensitive. It drafts, flags its guesses, asks you a few gap questions, and waits for your approval. Nothing is invented.

It's a short collaboration, not a single button — a couple of questions, one approval. After that, your context just loads. Manifest captures; your AI tool organizes and maintains.

Across your tools

Chat tools. Using the link? Nothing to do — an app with your connector starts every chat already knowing you. On GitHub? Paste the prompt that names your repo (a connected tool can see all your repos, so "my repo" won't do — the address is what picks the right one). And your manifest is plain markdown, so you can always export it as a file and attach it anywhere else.

Coding tools. Claude Code, Cursor, and Codex read the CLAUDE.md / AGENTS.md in whatever repo you open. If your Context OS is that repo, it's automatic. If not, drop a one-line pointer to it in a file the tool already reads, and it loads every session.

Load levels — only what a task needs

Stuffing your entire context into every prompt is the bloat this exists to kill. So MANIFEST.md sorts your context into levels, and a tool pulls only the smallest bundle the task actually needs:

Less spend per prompt, and the sensitive material stays out of the room until you put it there on purpose.

Keeping it current

At the end of a session, just tell the model "update my manifest." Through the link, your app saves it directly; in a connected repo, your tool does it with your approval.

One honest limit: a true "remind me before I close" depends on the tool. Manifest bakes a nudge into your context so models prompt you within a session, but a hard cross-tool reminder isn't something a portable file can enforce.

One source of truth

Over time, many models touch your context — Claude, ChatGPT, a coding agent, all of them. Each sees only a slice of you, so no version is right just because it came last. Context OS treats every tool as a partial witness, not an author.

So it converges instead of piling up contradictions. While a fact is still unsettled it carries its provenance — observed, reported, inferred, confirmed; once it's confirmed, it becomes plain current-state language and the scaffolding drops away. Settled facts stay settled. Your direct correction always wins, immediately. The same useful context emerges no matter which tool got there first.

Privacy

Everything you build on this page stays in your browser until you choose how to keep it. Choose GitHub — connected or do-it-yourself — and your context lives in your own private repo; no server of ours ever stores it, and the GitHub token lives only in your browser, scoped to the one repo. Choose the link, and we keep exactly one thing for you: your manifest, encrypted with a key made from your link — and the link itself is never stored, so it's the only key. Export or delete it anytime. The server that does this is open source — read exactly what it does, or run your own.

Whichever you choose, we gather nothing about you and sell nothing. Privacy is in the build, not a policy.

Questions

Why do I need Manifest?
Three reasons: consistency, privacy, and freedom.

Consistency

Anyone who's used AI past a simple chat knows a model's memory is limited, inconsistent, and sometimes just wrong. New chat — context lost. Switch models — most of it gone. Manifest gives every tool the same full picture of you and your projects, every time.

Privacy

Models work better the more they know — but no one wants an app holding everything about them: personal and professional, strengths and weaknesses, hopes and dreams. With Manifest you own your context, forever. You give a model only what the task in front of you needs, and nothing more.* You get the power of deep context without ever giving that power away.

* What a model retains on its own end varies — but nothing in your manifest is shared unless you choose to share it.

Cost

The subsidy era of AI is over. Token costs are skyrocketing — and being locked to one model leaves you at its mercy.

No model lock-in

Manifest lets you move to a stronger — or cheaper — model without re-explaining yourself to each new tool. Your go-bag is packed; don't like the price, or the model? Move on. When Fable 5 was suddenly pulled from the market by the U.S. government, anyone whose workflow depended on it was left stranded — unless their context was portable. As the landscape changes weekly, that portability is power: hop to the strongest model available, fast.
Won't all this context cost a fortune in tokens?
What costs a fortune is staying in one endless session because you're afraid to lose context, or burning tokens re-explaining the same thing for the tenth time. Manifest loads only the lightest context the task in front of you needs — that's how the Context OS is built. More relevant context, less spend.
What's the catch? Free tools this useful usually have one.
No catch. Manifest is free — every part of it, no gate, no trial. If it's useful and you feel like it, there's a suggested $5 donation — that's the whole business model. Keep an eye out for v2 and beyond.
It's free because you sell my data, right?
No. Manifest gathers nothing about you and sells nothing. With the GitHub options we never have your context at all — it lives in your own repo. With the link option we keep one thing: your manifest, encrypted, with your link as the only key — never read, never mined, deletable anytime. Privacy is the point. It's in the build.
Do I need to know how to use GitHub?
No. The link option needs no GitHub and no account — paste your context, get a link, add it to your apps once. The GitHub options are for people who want to hold everything themselves.
Which AI tools does this work with?
All of them. Apps that take connectors (Claude on every plan, ChatGPT on paid plans) read your manifest through your link. Repo-aware tools (Claude Code, Cursor, Codex) read it straight from your repo. And your manifest is plain markdown, so you can always export it as a file and attach it to anything else (Gemini, Perplexity). One context, every model — that's the whole point.
What exactly is a "Context OS"?
Your context (who you are, what you're working on, how you work, your goals) plus the operating layer that makes it work everywhere: load-levels so a model pulls only what a task needs, adapters so it behaves across Claude Code, Cursor, ChatGPT and the rest, a privacy model, and an update loop to keep it current. All plain markdown, human-readable, entirely yours — that operating layer is what makes it more than a text file. Manifest builds it for you.

About

Manifest (manifest.md) is a free tool for building a portable, private context for AI, and carrying it into every model you use. It started from a simple frustration: re-teaching every new AI everything you'd already taught the last one. Your context should belong to you, not to whatever app you used last. So Manifest gathers nothing about you and sells nothing — your context belongs to you, wherever you choose to keep it. Built in beta, free, and getting better — kept going by a suggested $5 donation from anyone it helps. Keep an eye out for v2.
manifest.md · readthemanifest.net · your context is yours