Private audio workspace
Turn recordings into useful work. Keep control of the source.
Transcribe, search, review, and produce recordings in one workspace. Self-hosted processing keeps media on your hardware by default; optional hosted services and external providers change that boundary.
Private beta. Docker required for self-hosting. CPU transcription; optional voice generation requires additional setup.
Concept illustration of a local audio workspace.
From recording to reviewed output.
Real product captures: local CPU transcription, source-linked notes and exports. Podcast, voice and automation workflows require separate evaluation.
Import
Upload a short recording to a running VoxNotch instance.
View full-size screenshot
Actual product capture: importing the garden-planning sample. Review
Read four timed segments produced by local CPU transcription and compare them with the source audio.
View full-size screenshot
Actual product capture: the completed sample transcript. Export
Download the transcript as Markdown, JSON, or SRT. Keep manually written source-linked notes in separate exports.
View full-size screenshot
Actual product capture: choosing a transcript export.
Start with the work you need to do.
Choose where processing happens.
Local defaults
Self-hosted transcription and media storage start inside your instance. Local sign-in is the default.
Explicit external services
Cloud models, remote storage, and outgoing webhooks are choices you configure. Hosted access puts media with the service operator.
Your operating responsibility
For self-hosting, you manage access, backups, updates, and retention. Installed engines may download model weights on first use; prepare them in advance for offline operation.
Know the requirements before evaluating.
Self-hosting requires Docker, storage, and ongoing maintenance. Transcription runs on CPU; speed and accuracy depend on your model, hardware, and recording.
Speech synthesis requires an optional engine and a CUDA GPU. The default mock engine produces a tone for testing, not spoken audio. Beta access and supported deployment options are agreed before evaluation.
Bring one workflow to the beta.
Tell us the recording type and output you need. Begin evaluation with a short, non-sensitive sample.
