Under the hood
How VidClean Works: Infrastructure & AI Models
VidClean is a free suite of 25 video and audio tools. This page documents exactly how it runs — the infrastructure, the AI models, and why it stays free — with specific, verifiable facts.
What AI models does VidClean use?
VidClean uses OpenAI Whisper (the "small" model, int8-quantized) for transcription and subtitles, and DeepFilterNet3 for background-noise removal, audio repair, and speech enhancement. Both models are self-hosted and run quantized on CPU, so there is no per-call API cost on the free tier.
How is VidClean's infrastructure set up?
VidClean runs on Railway for about $33/month across four services: a web service (the site and upload API), an FFmpeg worker for video and audio processing, a DeepFilterNet worker (8 GB RAM) for noise removal, and a Whisper worker (2 GB RAM) for transcription.
What happens to my files?
Files are processed entirely server-side and automatically deleted within 15 minutes. Uploads up to 2 GB are supported. Nothing is kept, shared, or used to train models.
Why is VidClean free?
Because the AI models are self-hosted and quantized to run on CPU, there is no per-use API cost to pass on. The free tier has no per-use cost to the user and no time limits. Total infrastructure is about $33/month, paid by the founder.
How many tools and how much has VidClean processed?
VidClean offers 25 free video and audio tools and has processed 581+ files to date, all free, with no account required and no watermark on any output.