Upload Workflow Only • English Speech • Runs 100% in Your Browser

English Instagram Video Transcript Generator

Upload an Instagram Reel or video file (.mp4, .mov, or .webm) that you have saved to your device. This tool extracts the audio track and transcribes spoken English into copyable text directly in your browser using an open-source Whisper speech model—with no account login and no paid API.

Saved video upload workflow: This version accepts saved video files (.mp4, .mov, .webm, or .m4a) from your phone or computer and does not support pasted URLs yet.

1. Upload Saved Instagram Video

Max 50 MB • Up to 3 min

Choose a saved Instagram video file or drag & drop it here

Accepted formats: MP4, MOV, WebM, M4A • Max file size: 50 MB • Max duration: 3 minutes (180s) • Spoken language: English

2. English Transcript

Words: 0 Characters: 0 Audio Duration: —

Tip: Once transcription finishes, you can switch between clean paragraph text and timestamped segments before copying.

How to Transcribe a Saved Instagram Video

Step 1

Save the Video on Your Device

Download your own Instagram Reel from your Instagram archive or camera roll, or use an MP4 video file you have permission to save on your phone, tablet, or computer.

Step 2

Select the MP4 / MOV File

Click Select Video File above and choose your video (up to 50 MB and 3 minutes long). Your browser verifies the video duration and extracts the audio track locally.

Step 3

Run Browser Speech-to-Text & Copy

Click Transcribe English Audio. On your first visit, your browser downloads and caches the open-source English Whisper model. When complete, click Copy Transcript.

Where Your Uploaded Video Is Processed (Privacy)

  • Zero Video or Audio Server Uploads: Your selected Instagram video file and decoded audio stay on your computer or phone. They are never uploaded to transcript.getswitchhelp.com or any other server.
  • In-Browser Audio Decoding: Your browser reads the file using the standard HTML5 File API and Web Audio API (AudioContext / OfflineAudioContext) to convert the audio track into 16 kHz mono PCM samples in device memory.
  • On-Device Open-Source Speech Model: Speech recognition runs inside a browser Web Worker using WebAssembly and the open-source Xenova/whisper-tiny.en model (OpenAI Whisper Tiny English, Apache-2.0 / MIT licensed).
  • One-Time Model Download & Browser Cache: The first time you transcribe a video, your browser downloads the quantized ONNX model weights from Hugging Face / jsDelivr CDN and stores them in your browser's local Cache Storage for subsequent runs.

Browser, Device & Model Limitations

English Speech Only

This tool uses the English-only whisper-tiny.en speech model. Videos with non-English speech will not be translated or accurately transcribed.

Saved Video Files Only (No Links)

This version accepts saved video files (.mp4, .mov, .webm, or .m4a) and does not support pasted URLs yet.

50 MB & 3-Minute Cap

Because audio decoding and speech recognition run in browser memory, uploads are limited to 50 MB and 180 seconds (3 minutes).

First-Time Model Download

The initial run downloads the open-source speech model files in your browser. Use Wi-Fi or a stable broadband connection for the first transcription; afterward the model loads from your browser cache.

Processing Time & Mobile Tabs

Processing time varies by device, video length, and first-time model download. Locking your phone screen or switching tabs can pause browser Web Workers.

Background Music & Noise

Clear spoken English produces the best results. Loud background music, heavy sound effects, or overlapping voices can cause missed or misheard words.