Voice Note Transcribeby SGraph — Simple. Private. Portable.
FIRST BETA — AUGUST 2026

Your WhatsApp voice notes, turned into a transcript, an analysis and a debrief — in one pass.

Drop in the audio. That's it. No accounts to configure, no steps to chain, no guessing whether transcription will work today. An infographic too, if you ask for one.

drop in the audio  (.opus / .ogg / .m4a)
         |
         v
ONE PASS  ->  transcript
          ->  analysis
          ->  short debrief
          ->  infographic (only if asked)
         |
         v
done.  no accounts to configure, no steps to chain.
Why this exists

General assistants can read your voice note. They just can't always transcribe it.

In our tests, a mainstream AI assistant identified a genuine WhatsApp voice note perfectly — codec, container, duration, sample rate — and still answered that it had no speech recognition available in that session. You can't predict when it will work. We do one job, and it works every time.

Made for WhatsApp audio

Voice notes are Opus audio in an Ogg container — and sometimes AAC when forwarded between iPhone apps. We accept both and detect the format from the file's content, never the extension.

One complete job

Not a step in someone else's workflow. Upload a voice note, get everything you need back at once: the words, the meaning, the summary — and a visual if you want it.

Nothing to install, nothing stored

Everything runs in your browser or app. There is no backend holding your audio — your files and your credits stay with you.

How it works

From voice memo to insights in seconds

The whole journey, end to end:

1

Upload

Drag in the voice note from WhatsApp — laptop download, Android export or iPhone forward.

2

Choose options

Pick your privacy mode, and whether you'd like an infographic with your results.

3

One pass

Transcript, analysis and a short debrief come back together. No chaining steps.

4

Export

Copy, download as markdown, or save the infographic. Your results are yours.

Privacy

Privacy is a mode you choose, not a disclaimer you scroll past

Different voice notes deserve different handling. Pick per upload:

ModeWhere it runsPrivacyPrice
ROUTED — DEFAULT Any provider via our model router No guarantee which provider processes it — we say so up front Cheapest
RESTRICTED Only providers you choose Named processors Higher
BROWSER-LOCAL Your own device Nothing leaves your device Possibly free
Honesty note: the default mode is the least private — requests are routed to whichever model provider serves them, and we cannot promise which one. If that matters for a recording, use restricted or browser-local mode. A full privacy notice ships with the beta.
Pricing

Pay for what you use. Around 20p per transcription.

No subscription. Buy credits, spend them when you need them.

Pay as you go

Each transcription priced at or under ~20p, passed through close to what the work actually costs.

Credits from £5

Top-ups start at £5 so card fees don't eat your credit — at £1, fixed fees would take 40% of it. We'd rather tell you that than hide it.

Your credit, your key

Your balance is provisioned as your own metered key with a hard spend cap. No invoices, no billing system, no surprises.

Early access opens with the August 2026 beta — capped, free-to-try credits for the first users. The browser-local mode may be entirely free.

Built in the open

Everything about this product is public

The code, the documentation, the design decisions — even the commercial arrangement behind it — are published as they are produced, under open licences (docs CC BY 4.0). Openness is the strategy, not a side effect.

View the repository