thewh1teagle/vibe

What it solves

Vibe provides a private, offline way to transcribe audio and video files into text. It eliminates the need to upload sensitive data to cloud services by processing everything locally on the user's device.

How it works

The application uses AI models like Whisper, Nemotron 3.5, and Parakeet TDT v3 to convert speech to text. It is optimized for various GPUs (Nvidia, AMD, Intel) across macOS, Windows, and Linux. Users can transcribe local files, system audio, microphone input, or audio from popular websites like YouTube and Vimeo.

Who it’s for

It is designed for anyone needing high-quality transcription services that prioritize privacy, as well as content creators who need subtitles for videos and reels.

Highlights

• Offline Privacy: Fully local transcription ensures no data leaves the device.
• Multilingual Support: Transcribes almost every language and can translate audio to English.
• comodule: Supports multiple export formats including SRT, VTT, TXT, and PDF.
• AI Analysis: Integrates with Claude API and Ollama for local AI-powered summaries of transcripts.
• Flexible Input: Supports batch transcription, system audio, microphone, and remote recording via phone QR code.
• Developer Tools: Includes a CLI and an HTTP API with Swagger documentation.