BUILD CAMPAIGNS THAT CONVERT

Transcriptr — Local Instagram Transcription
100% private. Your files and transcripts never leave this device. History is stored in your browser. For best results, bookmark and return from the same device and browser.
STEP 1 — INPUT WASM LOCAL
Drop your video or audio file here
or click to browse — no upload to server, ever
MP4 MOV M4A MP3 WEBM WAV
MODEL:
LANGUAGE:
Help & Testing Checklist
  1. Upload MP4/MOV/M4A → watch progress bars → transcript appears.
  2. Toggle timestamps on/off → export .srt/.vtt and verify captions.
  3. Open History → search, rename, delete, re-open any record.
  4. Switch model (tiny → base) and observe load time & accuracy.
  5. Verify no network calls in DevTools → Network tab (filter XHR/Fetch).
  6. If service worker is enabled: reload while offline to confirm cached app works.
Note: Instagram does not allow direct in-browser downloads (CORS/auth). Save the video from the Instagram app first, then upload here.
Model & Language
Smaller = faster; larger = more accurate
Auto-detect works for most content
Appearance
Light or dark interface
Storage
Calculating…
Privacy & Performance
All transcription runs in your browser
ACTIVE
SharedArrayBuffer / COOP-COEP detection
Detecting…

Usage Guide — Neural Architecture

1
Input: Paste Instagram link (optional) → Upload your saved video/audio file.
2
Decode: In-browser ffmpeg extracts and normalizes audio to 16 kHz mono WAV.
3
Chunk: Audio is segmented into manageable frames for efficient processing.
4
Encode: Whisper's WASM model converts audio frames into tokens.
5
Decode to text: Tokens become words with optional timestamps.
6
Review: Edit the transcript, toggle timestamps, add notes.
7
Export/Save: Copy or download (.txt / .srt / .vtt); history stays in your browser.