What we mean by 'personal' AI
The phrase gets used loosely. Here's the specific bar we hold ourselves to: your data, your retrieval, your choice of AI, your choice to walk away.
By BrainTube Team · · 5 min read
'Personal AI' has come to mean almost nothing. We use it for something specific: a system where you own the corpus, choose which AI reads it, can swap that AI without rebuilding anything, and can export everything in an afternoon. If any of those four is missing, the AI isn't personal — you're renting it.
You own the corpus
Your highlights, notes, brains and digests are yours. We store them, we don't train on them, and the export endpoint returns them in plain JSON whenever you want. Deleting your account deletes the corpus. There is no shadow copy we keep 'for analytics'.
You choose the AI
BrainTube does the capture, transcription, indexing and retrieval with its own routed mix of models — you don't manage keys or pick a model. What you do choose is which AI reads the result: connect the same library to Claude, Cursor, ChatGPT or any MCP-capable client. The corpus is the constant; the assistant on top is your call.
You can swap the AI without rebuilding
Because the corpus lives in one place and the retrieval interface is MCP, moving from one assistant to another is a repoint, not a migration. When a better client ships next month, you connect it to the same endpoint and carry on. You're never stuck with the app you started in.
You can walk away
We publish a one-page data pledge: what we store, how long, who can read it, and the exact export format. Walking away takes one click and one download. We'd rather lose users we can't keep happy than trap them.
Takeaways
- Personal AI = own corpus + choose the AI that reads it + swap it without rebuilding + clean exit.
- Anything less is rental, regardless of the marketing.
- Our data pledge is one page on purpose.
Try BrainTube on your own corpus
Free tier, no card. Export anytime.
More to read
- Second Brain vs AI Memory Layer: What Actually Changed — A second brain system organises what you write. An AI memory layer makes what you consume queryable by any model. The difference decides which one you should actually build.
- Stop watching videos. Start querying them. — The 60-minute video you bookmarked won't get watched. But the four answers buried in it will get used — if you can pull them on demand.
- How to build a second brain from your YouTube history — A practical workflow for turning years of passive watching into a knowledge graph you can actually use.