Free: Core transcription and review without file or time limits. Pro: US$64 one-time purchase for one licence covering three devices.
Free: $3 credit included with new accounts; Pay-as-you-go: $1/hour; Starter: $9/mo (unlimited, up to 2 hours per file); Pro: $19/mo (unlimited, up to 5 hours per file); Business: $49/mo (unlimited, up to 10 hours per file). Annual plans include 3 months free.
Key features
Edit transcripts with timestamps
Automatic speaker detection and naming
Local translation to 25 languages
AI research with custom prompts
Batch transcription of multiple files
Watch folders for automatic transcription
YouTube and Twitch direct transcription
Export as TXT, SRT, WebVTT, DOCX, PDF, Markdown, JSON, CSV
Custom vocabulary and automatic replacements
Voice memo recording and transcription
Speaker labels and word-level timestamps
52 languages and dialects
Chapters with timestamps
Search and Q&A with citations
Subtitle generation in multiple formats
Integration with ChatGPT, Claude, and AI agents
API and MCP server access
Encrypted storage and privacy controls
Library search across multiple recordings
File upload for meetings, lectures, podcasts and voice notes
What makes it different
Local processing on your machine with no account or upload required
One-time Pro purchase with no subscription
Works offline after models are downloaded once
Runs on CPU and Vulkan without requiring powerful graphics card
Audio stays private on device, not sent to servers
Word-level timestamps that enable precise subtitle generation for any platform
Ask questions across multiple recordings with citations back to the original moment
Supports use directly within ChatGPT and Claude without leaving your agent
Benchmarked accuracy with 1.5x to 2.3x fewer errors than Whisper large-v3
Permanently searchable library that organizes all transcripts in one place
TrueScribe
Transcribe.so