Scriberr, the open source self hosted audio transcription app: upload audio and get accurate whisper powered transcripts in a clean web interface, secured from first boot by cloudimg.
Scriberr is an open source, self hosted AI audio transcription application. You upload an audio recording and it returns an accurate, searchable transcript through a clean web interface, powered by OpenAI Whisper models via WhisperX. It can optionally separate speakers with diarization and summarise or let you chat over a transcript with your own language model, and everything stays on your own machine so recordings never leave your infrastructure.
It suits anyone who needs private transcription at scale, journalists, researchers, podcasters, support and legal teams, who want their recordings turned into text inside their own cloud account rather than sent to a third party service.
The cloudimg image ships Scriberr fully installed behind an nginx reverse proxy with the WhisperX transcription environment and a compact Whisper model pre baked, so a transcription works immediately and fully offline the moment you log in, with no multi gigabyte model download on first use. No administrator ships in the image: a unique admin password is generated for your instance on first boot, written to a root only file, and self registration is then closed so no one else can create an account. The per instance session signing secret is generated uniquely on first boot, the default profile is tuned to run on a modest CPU only VM, and the paired deployment guide plus 24/7 support cover logging in, transcribing, and enabling diarization and summaries.
Real screenshots taken while testing this image against its deployment guide.