PhotoSplit Studio
Drop your audio or video file here
or click to browse
MP3
MP4
WAV
M4A
OGG
WEBM
MOV
FLAC
—
⚠️
or
Preparing…
📄 Transcript
0 words
🔒
Privacy: Your audio is processed
entirely locally inside your browser using
WebAssembly and WebGPU. Audio files and transcripts are never
uploaded to any server. PhotoSplit Studio does not store, log,
or access your data.
How It Works
1
Local AI Model
Your browser downloads a compressed Whisper AI model (one-time download, cached locally).
2
Upload or Record
Drop a meeting recording (MP3, MP4, WAV…) or record live with your microphone.
3
Transcribe Offline
The AI model runs directly on your device's hardware to generate the text privately.
4
Export
Download as plain text, Markdown report, Word document, or print to PDF.
Frequently Asked Questions
Is this tool free and private?
Yes, 100% free and 100% private. Unlike other services, we do
not upload your audio to the cloud. The transcription happens
completely on your device using your CPU or GPU.
Why does it say "Loading Model"?
The first time you use the tool, your browser downloads a
compressed version of the AI model (~150MB for Tiny, ~300MB
for Base). It is saved in your browser's cache so it loads
instantly next time.
What audio formats are supported?
MP3, MP4, M4A, WAV, OGG, WebM, MOV, and FLAC — up to 50 MB per
file. Because the processing happens in your browser,
extremely long files might cause your browser to slow down.
How accurate is the transcription?
We use an optimized version of the famous OpenAI Whisper
model. The "Base" model is highly accurate for clear speech.
The "Tiny" model is faster but might make more mistakes with
heavy accents or background noise.
Can it identify who is speaking (diarization)?
The current browser-based Whisper models do not support
speaker identification. If you need "who said what" labels,
you'll currently need an external service.
What output formats can I download?
Plain text (.txt) — the raw transcript.
Markdown (.md) — a structured report with metadata, word count, and date.
Word / RTF — opens in Microsoft Word, Google Docs, LibreOffice.
Print / PDF — opens a print dialog so you can save as PDF. You can also edit the transcript directly in the text area before exporting.
Markdown (.md) — a structured report with metadata, word count, and date.
Word / RTF — opens in Microsoft Word, Google Docs, LibreOffice.
Print / PDF — opens a print dialog so you can save as PDF. You can also edit the transcript directly in the text area before exporting.
Does live recording work in all browsers?
Live microphone transcription uses the Web Speech API, which
is supported in
Chrome, Edge, and Safari. It is not supported
in Firefox. For file-based transcription (upload), all modern
browsers are supported.