Auphonic

Auphonic's site doesn't tell you who built it. There's no founder letter on the homepage, no origin story and no team page in the material we reviewed. What it does tell you is why the product exists: it calls itself "your AI sound engineer for podcasts, education, videos, audiobooks, radio and broadcasts." That's a narrow, practical goal. It doesn't promise to write your show or clone your voice. It promises to fix the audio you already recorded, and it claims 2M+ users and "leading audio companies" as proof that this approach works.

The product is built around that sound engineer idea. You drop in an audio or video file, pick an algorithm or preset, and Auphonic handles the post production steps a human engineer would. Options include Voice Cleaner (with or without music), a Static/Music Denoiser, partial dereverb, an Adaptive Leveler that balances speakers and music "no compressor knowledge required," De-Esser and De-Plosive filtering, clipping repair and bandwidth extension. It also cuts filler words like "uhm" and "ähm" in multiple languages, removes coughs and silence, and hits loudness targets with true peak and MaxLRA settings. Multitrack processing adds automatic ducking, a noise gate and mic bleed removal. The part that matters most for working creators is that you set a preset once and stop thinking about levels.

The recent changelog shows the product is still being actively worked on: an n8n integration, direct multichannel file import, an editable gain curve for refining the leveler, burned-in styled subtitles and a Studio Voice beta. Around the audio engine sit watch folders, SFTP, Dropbox and Google Drive transfers, automatic publishing to platforms like YouTube, and a public API. Pricing is freemium: you get 2 hours of processed audio free each month, and paid recurring credits start at $11/mo, with 20% off if you pay yearly. If you publish regularly and hate mixing, that's an easy trial to justify.