Observed arrival · 2026-10-20
Soundpoint: Russian speech, word by word
A Russian transcription service that links words to playback and describes tools for proofreading, speaker labels, audio cleanup, and exporting.
- For
- People transcribing long Russian-language recordings
- Worth noticing
- The page names Sber’s GigaAM v3 and NVIDIA’s TitaNet, and says audio does not go to third-party services.
Field notes
The homepage lists MP3, M4A, WAV, MP4, and MOV inputs, plus TXT, DOCX, SRT, VTT, CSV, MP3, and WAV outputs. It describes speaker separation with user-assigned names, whole-word find-and-replace, chapters, and A/B comparison of audio processing on a selected segment. The page identifies Sber’s GigaAM v3 for transcription and NVIDIA’s TitaNet for speaker differentiation; its privacy and model-performance statements are presented by the service, not independently verified.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue