SolScribe
Self-hosted audio transcription powered by WhisperX. Record, transcribe, search, and chat with your audio, all on your own hardware. Your data never leaves your network.
Three Steps to Searchable Audio
Record or Upload
Record directly in your browser or upload audio files. Supports WAV, MP3, M4A, FLAC, and more.
Transcribe with WhisperX
WhisperX processes your audio locally with word-level timestamps, confidence scores, and speaker labels.
Search, Chat & Export
Full-text search across all transcripts. Chat with an LLM about your audio. Export as SRT, VTT, TXT, or JSON.
Everything You Need for Audio Intelligence
WhisperX Transcription
State-of-the-art speech recognition running locally on your hardware. Fast, accurate, and completely private.
Speaker Diarization
Automatically identify who said what with PyAnnote-powered speaker segmentation. Label and track every voice.
LLM Chat
Ask questions about your transcripts in natural language. Get summaries, extract action items, or find specific details.
AI-Powered Analysis
Auto-generate summaries, key discussion points, decisions made, and follow-up action items from any transcript.
Confidence Highlighting
Word-level confidence scores let you see exactly which words the model was unsure about. Color-coded for quick review.
Full-Text Search
Instantly find any word or phrase across your entire transcript library. Jump directly to the moment in the audio.
Webhook Automation
Trigger n8n workflows, Zapier zaps, or any HTTP endpoint when a transcription completes. Automate your pipeline.
100% Self-Hosted
Your audio and transcripts never leave your network. Run on your own server with Docker, CPU or CUDA GPU.
Simple, One-Time Licensing
1 user, all core features, self-hosted
Unlimited users, API access, RBAC, 500 cloud minutes
BAA included, audit logs, encryption at rest, compliance documentation
One-time purchase. Cloud compute credits available separately.
Built with
Runs on any Linux server with Docker. GPU acceleration optional but recommended for faster transcription.
Get Started in 5 Minutes
Clone the repo, run Docker Compose, and start transcribing. No API keys, no sign-ups, no data leaving your machine.
docker compose up -d