Hinglish reel captions, made offline on your own laptop

Description and use

A local web app that uses Whisper to transcribe Hinglish audio, converts Devanagari to Roman script, and lets editors edit captions while preserving timing, enabling cost‑effective, privacy‑preserving captioning.

Problem

Freelance editors and small agencies struggle to produce Hinglish captions that match the spelling users actually type, while keeping timing intact and avoiding cloud costs.

Outcome

The tool reduces captioning time by automating transcription and Romanization, keeps timing stable during edits, and eliminates cloud costs and data transfer, enabling freelancers to scale their side‑hustle without additional per‑minute fees.

Workflow

• Drop video or audio into the web interface
• Whisper transcribes and outputs timestamps
• App converts Devanagari Hindi to Roman Hinglish using custom rules
• Captions are displayed word‑by‑word in the preview
• Editor allows spelling fixes, splits/join words while keeping timing
• Choose style or brand colours
• Export as burned‑in MP4, SRT, VTT, ASS, or plain transcript
• Optionally run a local LLM (Ollama) to tidy spellings

Evidence

The problem is not that speech recognition can't hear Hindi, because open models like Whisper hear it well. The problem is what comes out.

Source

Hinglish Reel Captioner is a small local web app for that job. You drop in a video or a voice note. It listens with open-weight Whisper, converts each Hindi word to Roman Hinglish while leaving English words English, and plays big word-by-word captions over a live preview.

Source

The editor keeps the timing intact, and the optional LLM pass follows the same discipline.

Source