Scriberr
Scriberr is a self-hosted, privacy-focused AI transcription app for audio and video, with speaker diarization, word-level timestamps, summaries, and transcript chat.

Scriberr is an open-source application for transcribing audio and video locally, designed to keep recordings private by avoiding third-party cloud processing. It provides a web-based interface to upload, record, review, and work with transcripts, with optional integration for LLM-powered transcript chat and summaries.
Key Features
- Local/offline transcription using modern speech-to-text models (including Whisper and newer model options)
- Speaker diarization to separate and label different speakers
- Word-level timestamps and transcript playback follow-along with seeking from text
- Built-in audio recorder plus note-taking/annotation while listening
- Transcript summarization and “chat with your audio” (supports local LLMs via Ollama and OpenAI-compatible providers)
- Automation-friendly features such as an API and folder watcher for auto-processing new files
- PWA support for a more native app-like experience on desktop and mobile
Use Cases
- Transcribe meetings, interviews, and lectures without uploading sensitive audio to external services
- Process large batches of recordings automatically via folder watching and API-driven workflows
- Create searchable, annotated transcripts and generate summaries for personal knowledge capture
Limitations and Considerations
- High-accuracy transcription and diarization can be resource-intensive; GPU acceleration is recommended for best performance
- Some advanced features (like transcript chat) may require configuring external or local LLM providers
Scriberr is a strong fit for privacy-conscious users who want reliable local transcription with a polished review experience and workflow automation options. It combines transcription, organization, and AI-assisted analysis into a single self-hostable service.
