Scrieb
Offline transcription, speaker labels and summaries for Mac and Windows
Overview
Scrieb records or imports audio and video and turns it into a transcript, a summary and a searchable library. It is built for people who transcribe interviews, meetings, lectures and voice memos but can't or don't want to upload them: journalists, researchers, lawyers, therapists and teams under GDPR. It runs on macOS and Windows, with a free plan of one hour a day. All AI runs on the user's computer. OpenAI's Whisper (Small, Medium or Large v3 Turbo) transcribes speech in 90+ languages, a local model separates speakers, and a local language model writes summaries by type (interview, meeting, lecture or voice memo) and answers questions in a chat across all recordings. No audio, transcript or summary is sent to a server, and an offline mode turns off every connection.
Pricing: $24.99
Minimum macOS: 14.0 (Sonoma)
Architecture: Apple Silicon
Key Features
- Fully offline transcription with Whisper Small, Medium or Large v3 Turbo
- Transcription in 90+ languages, with automatic language detection
- Speaker identification, including "Me" versus other participants
- Local AI summaries: interview, meeting, lecture, voice memo or custom instructions
- AI chat across all recordings, naming the recordings it used
- Record from the microphone or from another app, such as Zoom or Teams
- Import audio and video: MP3, WAV, M4A, FLAC, MP4, MOV, WebM and more
- Search inside all transcripts and summaries
- Custom vocabulary for names, product terms and acronyms
- Export to DOCX, PDF, TXT and Markdown
- Offline mode with no automatic network connections
- macOS (Apple Silicon and Intel) and Windows 10/11 (x64 and ARM64)
- Free plan: 1 hour a day, no card needed
Scrieb Alternatives
MacWhisper
The all-in-one Mac app to transcribe it all
OpenWhispr
Open-source voice-to-text dictation with local and cloud processing
Whisper Transcript
On-device speech-to-text powered by Apple Foundation Models
AudioWhisper
Menu bar voice-to-text with multiple transcription engines and semantic correction