Self-Hosted Transcription Platform

Your Audio.
Your Hardware.
Your Rules.

WhisperX-powered transcription with speaker diarization, AI analysis, and confidence highlighting. Runs entirely on your network — no data ever leaves.

Get Started Free Explore Features
100% Self-Hosted
GPU Accelerated
MIT Licensed
How It Works

Three steps. Five minutes.

From raw audio to searchable, AI-analyzed transcripts. No cloud accounts, no API keys, no monthly bills.

01

Record or Upload

Record directly in the browser or upload any audio file. MP3, WAV, M4A, FLAC, OGG — all supported.

02

Transcribe Locally

WhisperX processes audio on your hardware with word-level timestamps, confidence scores, and speaker diarization.

03

Search, Chat & Export

Full-text search, LLM chat with your transcripts, auto-generated reports, and export in SRT, VTT, TXT, JSON, or HTML.

Capabilities

Everything you need.
Nothing you don't.

No feature bloat. No subscription tiers that lock core functionality. Every feature ships in the open-source community edition.

WhisperX Transcription

State-of-the-art speech recognition with forced alignment for precise word-level timestamps. Runs locally on CPU or CUDA GPU.

Speaker Diarization

Automatically identify who said what with PyAnnote-powered speaker detection. Label and rename speakers for clean transcripts.

LLM Chat

Ask questions about your transcripts in natural language. Uses any OpenAI-compatible API — local (LM Studio, Ollama) or cloud.

Auto-Export Reports

After every transcription: AI-generated summary, key points, decisions, and follow-ups — plus the full transcript with confidence highlighting.

Full-Text Search

SQLite FTS5 full-text search across every transcript. Find any word or phrase instantly, jump to the exact timestamp.

Confidence Scoring

Every word gets a confidence score from WhisperX. Low-confidence words are highlighted in the transcript so you know exactly what to double-check.

Webhook Automation

Fire webhooks on transcription complete, report generated, and more. Perfect for n8n, Zapier, or any automation pipeline.

100% Self-Hosted

Your audio files, transcripts, and analysis stay on your hardware. No cloud dependency. No data exfiltration. Guaranteed by architecture.

Auto-Export Reports

Every transcription generates an intelligence report.

SolScribe automatically analyzes your transcript and produces a structured report with AI-powered insights and word-level confidence highlighting.

  • Executive summary capturing purpose and outcome
  • Key points, decisions, and action items extracted
  • Full transcript with confidence-highlighted words
  • Export as HTML (printable) or Markdown
  • Shareable via unique URL or file download
meeting-standup-2026-03-14.html
Summary
Key Points
Transcript
00:00:02  Speaker A
Good morning everyone, let's go through the sprint updates. We shipped the new authentication flow yesterday and Miroslav is working on the API.

00:00:14  Speaker B
The PostgreSQL migration is done. I need to run the benchmarks before we can deploy to staging on Friday.
High
Medium
Low
API & Integrations

Built for automation. Not just transcription.

A full REST API, real-time webhooks, and first-class n8n integration. SolScribe fits into your existing workflows — not the other way around.

REST API

70+ endpoints covering every operation. Upload audio, trigger transcription, search across transcripts, chat with content, and manage webhooks — all programmatically.

API Endpoints
# Transcribe audio POST /api/v1/transcriptions/quick # Search across all transcripts GET /api/v1/search?q=budget+review # Chat with a transcript POST /api/v1/transcriptions/{id}/chat # Generate analysis report POST /api/v1/transcriptions/{id}/report/generate # Export as SRT/VTT/TXT/JSON GET /api/v1/transcriptions/{id}/export?format=srt
  • GET List, search, filter transcriptions
  • POST Transcribe, summarize, score, chat
  • DEL Clean up transcriptions & audio

Webhooks & n8n

Fire webhooks on every event: transcription complete, report generated, summary ready, score calculated. Build end-to-end automation pipelines with n8n, Zapier, or any HTTP endpoint.

Webhook Payload
{ "event": "transcription.completed", "transcription_id": "a1b2c3d4", "title": "Weekly Standup", "duration": 847, "word_count": 2340, "speakers": ["Alice", "Bob", "Carol"], "confidence_avg": 0.94 }

Example n8n workflow:

SolScribe
→
Webhook
→
LLM Summary
→
Email

Transcription completes → webhook fires → n8n summarizes with LLM → emails the team. Fully automated, zero human intervention.

Pricing

Start free. Scale when ready.

The community edition includes every feature. Team and cloud tiers add multi-user management and hosted convenience.

Community
Free
Open source, self-hosted, full-featured.
  • Unlimited transcriptions
  • All AI features
  • All export formats
  • Webhook automation
  • Community support
Get Started
Team
$29/mo
Multi-user with admin controls.
  • Everything in Community
  • Multi-user accounts
  • Admin panel & RBAC
  • Shared transcript library
  • Priority email support
Contact Us
Cloud Starter
$9/mo
Hosted for you. 5 hours/month.
  • Hosted infrastructure
  • 5 hours of transcription
  • All AI features
  • Auto backups
  • Email support
Coming Soon
Cloud Pro
$29/mo
Unlimited hosted transcription.
  • Hosted infrastructure
  • Unlimited transcription
  • Priority GPU queue
  • Custom retention policies
  • Priority support + SLA
Coming Soon
Get Started

Five minutes to your first transcription.

Docker Compose. One command. No API keys, no sign-ups, no data leaving your machine.

Click to copy # start SolScribe
docker compose up -d
View on GitHub
Go 1.21+ React 18 WhisperX 3.3 Docker SQLite GORM PyAnnote CUDA