Every AI has memory now. None of it is yours.

By 2026, memory stopped being a feature and became table stakes. ChatGPT remembers you by default. Claude and Gemini carry context across sessions. There are notes-vault apps that remember what you type, OS-recall tools that scrape your screen, and wearables that listen all day.

There is one thing almost none of them give you: ownership.

ChatGPT's memory lives in OpenAI. Claude's lives in Anthropic. Your notes app remembers only what you typed into that app. And the tools that promised to remember *everything* have a habit of disappearing — Rewind and Limitless were acquired by Meta in late 2025, and the Rewind desktop app was shut off that December. Memory you don't own is memory someone else can turn off, lock in, or walk away from.

That's the problem My Brain is built to solve.

What My Brain is

My Brain is a brain you own that follows you — across every AI tool, every session, every team, and every job. It is keyed to *you*, not to a vendor's model or a single app. When you switch from Claude to Cursor to ChatGPT, or leave one company for the next, the brain comes with you.

Two things make it different from the memory you already have:

is built on the same knowledge fabric RepoOps already keeps for your repos — your accumulated lessons, decisions, notes, and preferences — plus anything you capture. It's real working memory, not a chat log.

1. It's grounded in what you actually do, not just what you type. My Brain

cited answer drawn only from your own knowledge, with links back to the source. Not a model's fuzzy recollection of you — grounded retrieval over files you can open.

1. It reasons, and it shows its work. Ask it a question and you get a

And it's safe by construction: every write lands as a proposal you review, every memory carries its provenance, nothing is silently rewritten, and every external read is scoped, revocable, and audited.

How you use it — five ways in

My Brain isn't a single app you have to live inside. It's a brain with doors on every side.

1. Ask it, right on your desktop. Open the RepoOps dashboard, go to My Brain, and ask a question in plain language. You get a cited answer reasoned over your own notes — with a deep mode that does multi-hop reasoning for the hard questions. Runs locally; bring your own model key.

2. From Claude Desktop, Cursor, ChatGPT, or Cowork — via MCP. The Model Context Protocol is becoming the USB port for AI: one standard that Claude, GPT, and Gemini clients all speak. Connect My Brain once, and every MCP-capable tool you use can read *your* brain — ask_brain for a cited answer, search_brain for raw chunks, get_brain_file to pull a note. One connect; every assistant you touch now knows what you know.

3. From any app you build — with a scoped read token. Mint a minimally scoped, revocable read token and drop our copy-paste snippet into a website, a CRM widget, or a Google Sheet (=REPOOPS_ASK("…") via the Apps Script template). Now a real second property reads a slice of your brain over a simple REST call — the token stays server-side, header-only, and you can revoke it the moment you want to.

4. In Obsidian. Your brain is flat markdown with wiki-links, which means it *is* a valid Obsidian vault. Point Obsidian at it, install Dataview, and you get a graph view and live dashboards over your knowledge — and edits flow back through the same trust guard. Markdown doesn't rot.

5. Share a slice. Publish a curated slice of your brain as a brain-pack others can install. When someone installs it, it lands in *their* brain as proposals to review — with lineage tracking exactly where each idea came from. Knowledge that compounds between people, not just within one.

Across your whole life, not just your code

RepoOps started in the repo, but a person's brain isn't only about code. The same store holds the decision you made and why, the preference you want every assistant to respect, the lesson you learned the hard way, the thing you jotted at 11pm so you wouldn't lose it. Work and personal, in one owned place.

And it works *for* you while you sleep: a nightly consolidation pass — Dreaming — reviews what came in, surfaces connections, and drafts a morning brief. The captures flow in safely as proposals; you decide what becomes memory.

The point isn't another silo. It's the opposite: one brain that every tool in your life can reach, that you actually own, that gets more useful the longer you use it.

Where this is going: a brain you can just *ask*

Everything above already treats your brain as something you retrieve from and reason over. The natural next step is a single, searchable front end for your whole life — one place to ask *"what did I decide about X,"* *"what do I know about this person,"* *"pull everything I've learned about Y"* — across every source you've connected, answered with citations, from any device.

That's the direction we're most excited about: your brain, as a thing you can query as easily as you search the web — but private, owned, and about *you*. Call it your personal search engine for everything you know.

Own the thing that knows you

The tools that remember you today are betting you'll never leave, because your memory is trapped where they keep it. My Brain is the other bet: that the most valuable context in your life — what you've learned, decided, and built — should belong to you, follow you everywhere, and answer when you ask.

Free and local to start: repoops.ai

---

Sources:

• Meta acquires Limitless; Rewind app disables capture Dec 19, 2025 (TechCrunch, 2025-12-05): techcrunch.com/2025/12/05/meta-acquires-ai-device-startup-limitless/
• Rewind Mac app shutting down following the Meta acquisition (9to5Mac, 2025-12-05): 9to5mac.com/2025/12/05/rewind-limitless-meta-acquisition/
• Model Context Protocol — the open standard for connecting AI clients to tools/data: modelcontextprotocol.io