Features
Everything you need to transcribe audio in batches
LoheVox is made for teams that process lots of recorded audio and need reliable, well-formatted output.
Accurate, self-hosted speech models
We run modern open speech-recognition models on our own GPU infrastructure, tuned for long-form audio. We don't resell anyone else's transcription service as our product.
Smart chunking
Long recordings are cut at natural pauses using voice-activity detection, so sentences aren't split mid-word, and transcribed in parallel for fast turnaround.
Output formats
- TXT: clean paragraphs for reading and search
- SRT / VTT: subtitles ready for video players
- JSON: segments, word timestamps, confidence, detected language
Languages
90+ languages with automatic detection. Pass a language hint when you know it, for better accuracy and speed.
Developer-friendly API
A REST API with signed upload and download URLs, idempotent job creation, and HMAC-signed webhooks. See the developer overview.
Privacy controls
Encrypted at rest and in transit, short default retention, and deletion on demand. Read security & data.
Great for
Media & podcasts
Show notes, captions and searchable archives.
Research & interviews
Qualitative interviews and focus groups, ready for analysis.
Education
Lecture captions and accessible course material.