Dictation is free. Everything else is one payment.

Because MemoFlow costs almost nothing to run. Transcription, summaries and search all execute on your own Mac, so there is no per-user server bill that has to be covered every month. A subscription would be charging rent for compute you are already paying for.

Cloud notetakers price monthly because they genuinely have a monthly cost — every minute of your audio is uploaded, transcribed and stored on someone else's hardware. MemoFlow never uploads any of it, which removes that cost entirely and, more to the point, removes the reason your recordings would ever sit on a server in the first place.

It is sustainable for the same reason it is cheap: no servers to fund. The ongoing cost of MemoFlow is development time and an Apple developer account, not infrastructure that scales with how much you talk.

It is a fair thing to ask — cheap lifetime deals have a reputation, and usually the reason is that the maths never worked. Here the maths is unusually simple. A free dictation user costs nothing to serve, so the free tier can stay free without being subsidised by anyone, and a one-time payment goes toward continued work on the app rather than covering last month's inference bill.

MemoFlow is built by one developer. Paying once is what keeps that possible.

Nothing. There is nothing to stop paying. You buy once, and your recordings, transcripts and summaries are files on your own disk that stay readable whether or not MemoFlow is still installed.

This is the practical difference between local software and a service. A subscription that lapses usually takes your archive with it, or holds it behind a re-subscribe wall. Your MemoFlow library lives in ~/Library/Application Support/MemoFlow and belongs to you.

Unlimited system-wide dictation, on-device, in English, Hindi and Hinglish, with no account and no word limit. It is the whole dictation product, not a sample of it.

• —No time limit and no monthly word cap.
• —No account or sign-in — the app has no login at all.
• —Works with no network connection once a model has downloaded.
• —Custom dictation modes, word replacements and per-app behaviour included.