# transcriber

> Self-hosted, open-source web app for transcribing audio/video with Whisper. Supports YouTube, local files, real-time progress, dark mode.

- **Category:** Audio Processing
- **Licence:** MIT
- **Open source:** oss
- **Languages:** Python, HTML
- **Platforms:** Docker, Python
- **Tags:** Privacy, Machine Learning, Videos, Speech Recognition, Audio, Youtube, Gpu, Docker, Open Source, Fastapi, Transcription, Self Hosted, Whisper, Speech To Text, Audio transcription, Whisper Ai, Video Transcription, Youtube Transcriber

## Description

Transcriber is a privacy-first, self-hosted web application that transcribes audio and video files using OpenAI Whisper models. It accepts YouTube URLs or local file uploads (MP3, MP4, WAV, and more), processes them with either openai-whisper for accuracy or faster-whisper for speed, and streams real-time progress updates via WebSockets. The app includes a smart library that correlates media files with their transcripts, automatic filename sanitization, GPU acceleration with CUDA support and CPU fallback, and automatic FFmpeg installation. It runs via Docker (multi-stage, non-root) or locally with Python 3.10–3.12, using FastAPI, Alpine.js, and Tailwind CSS. The project is MIT-licensed with no paid tiers or cloud dependencies.

## Links

- [Homepage](https://github.com/BrendownFerreira/transcriber)
- [Website](https://noktech.com.br/softwares/transcriber)
- [GitHub Repository](https://github.com/BrendownFerreira/transcriber)
- [Issues](https://github.com/BrendownFerreira/transcriber/issues)
- [Pull Requests](https://github.com/BrendownFerreira/transcriber/pulls)

---
Source: https://selfhost.directory/project/transcriber