Cleanvoice is an AI podcast and video post-production tool that automatically strips filler words, stutters, mouth sounds, background noise, and dead air out of raw recordings, then returns a studio-cleaned file plus a transcript and show notes. It targets podcasters, video creators, and agencies who don't want to manually edit every "um" and lip-smack out of an episode. Pricing is usage-based and billed per hour of processed audio, with a 30-minute free trial and hourly rates that fall from roughly $2.20 to $0.90 as volume grows.
Who builds it
Cleanvoice AI was founded in 2022 by solo founder Adrian Spataru, who started podcasting in 2020 and grew frustrated that existing editing tools handled non-English accents poorly. The company is based in Bucharest, Romania. It says it serves 15,000+ podcasters directly, plus 30+ established brands that integrate its editing engine through the API for high-volume production.
Core features
- Filler word remover — cuts "um," "uh," "like" and similar fillers; officially trained for English, French, German, Romanian and Arabic, with reasonable results on related languages and accents such as Irish, Australian and German English
- Background noise remover — strips traffic, barking dogs, doors, wind, static and other ambient sound; this and the silence/mouth-sound tools work on any language since they analyze audio patterns, not speech content
- Silence and dead-air remover, mouth-sound/breath remover, and stutter remover
- Studio Sound audio enhancer — balances levels and removes echo/distortion for a "studio-quality" master
- Automatic transcription plus AI-generated podcast summaries and show notes
- Multitrack support for multi-guest recordings, and a REST API (docs.cleanvoice.ai) with Python/JavaScript SDKs and Make.com/n8n integrations
Pricing
- Free trial: 30 minutes of audio/video processing, no credit card required
- Pay-as-you-go credits (valid 2 years): $11 for 5 hours (~$2.20/hour), $20 for 10 hours ($2.00/hour), $45 for 30 hours ($1.50/hour)
- Subscriptions, unused credits roll over up to 3x the plan limit: 10 hours/month for $11 (~$1.10/hour), 30 hours/month for $30 ($1.00/hour), 100 hours/month for $90 ($0.90/hour)
- Custom/Enterprise plan for 200+ hours/month with dedicated API endpoints and priority support
- Minimum billing unit is 1 minute of processed audio; check the current pricing page as rates can change
Who it's for
Cleanvoice fits solo podcasters and small shows that just need the 30-minute free trial or a low-volume pay-as-you-go pack, and it scales to networks and agencies editing dozens of episodes a month via the 100-hour subscription or a custom API plan. Its billing model rewards regular producers: a weekly show that processes a few hours a month is cheapest on a subscription tier, while someone editing a handful of one-off interviews may prefer the pay-as-you-go credits that stay valid for two years. It's less suited to anyone needing filler-word detection in languages outside its officially supported set, or to pure transcription-only workflows where a dedicated speech-to-text API would be cheaper. Files are kept for 7 days after processing and the service is GDPR-compliant with EU data storage, which matters for European studios and agencies handling client audio under contract.