Trusted by creators at startups, agencies, and Fortune 500 teams.
Free forever, no signup, no watermark.
How do I fix bad audio in a screen recording?
Upload it above. The browser analyzes the audio and tells you whether it's too quiet, clipping, missing on one channel, or just uneven, then recommends one of five fixes. If you'd rather test your mic before recording again instead of fixing it after the fact, run the microphone test or the full recording readiness test first.
Avoid the fix-it-later cycle
Recording with Clipy captures audio the same way every time, so gain, channel routing, and levels stay consistent recording to recording — fewer trips back to a fixer tool after the fact.
How the in-browser analysis works
The browser decodes the audio track with the Web Audio API's decodeAudioData, then measures peak amplitude, RMS (an approximation of perceived loudness), the fraction of samples hitting the ceiling (clipping), and whether one channel is silent while the other isn't. None of this is uploaded — it's a local decode-and-measure pass that only produces a diagnosis and a suggested mode. Files over 100 MB skip this step to avoid a slow in-browser decode; the actual fix always happens server-side regardless of file size.
Why five separate fix modes instead of one 'auto-fix' button
Quiet audio, clipped audio, one-sided audio, and background noise all need different ffmpeg filters, and applying the wrong one can make things worse — boosting an already-clipped track adds more distortion, and denoising a merely-quiet track won't fix the volume. The analysis recommends a mode, but you can override it: sometimes a recording needs two passes (declip, then normalize), and running one fix per pass gives you a chance to check the result before combining fixes.
