Local, private AI transcription for meetings, dictation, and recordings that runs entirely on your computer
Convert PDFs into chapter-aware audiobooks with smart chapter detection
Stage
Not given
Startup
Founded
Not given
2024
Based in
GB
Not given
Pricing model
Freemium
Freemium
Pricing
$9 USD/month, $89 USD/year (17% discount), or $189 USD one-time lifetime license. All plans include 10-hour free trial with no account or credit card required.
Free: 5,000 words on signup plus 500 words monthly. Creator: $9/month with 50,000 words. Pro: $45/month with 200,000 words. Studio: $149/month with 600,000 words. Additional words available at $0.40 per 1,000 words.
Key features
On-device audio recording and transcription
Real-time speaker identification with cross-meeting recognition
AI-generated summaries, action points, and chapter markers
Import and transcribe audio and video files locally
System-wide dictation tool with optional LLM cleanup
Calendar integration with Apple Calendar and Google Calendar
Full-text search across all meetings
Export to Markdown, PDF, and via webhooks or MCP
Works fully offline with no internet connection required
Convert PDF to audiobook
Convert DOCX to audiobook
Convert EPUB to audiobook
Automatic chapter detection
Front matter and table of contents skipping
Chapter-grounded Q&A
Multiple narration voices
Resume playback across devices
MP3 and WAV downloads
What makes it different
No cloud-based servers or analytics; full ownership of your data with lifetime license option
Works completely offline with no account, login, or credit card required for trial
Integrates with your existing tools via MCP, webhooks, and direct API access without data intermediaries
Optional cloud LLM support using your own API keys, giving you control over which provider processes summaries
Chapter-aware audiobook format rather than continuous narration
Automatic detection of chapter structure and front matter classification
Private storage with no training or redistribution of uploaded content
Detects and surfaces figures, code, and equations as source pages during playback
talat
Audionarrator