Restore Old & Archival Recordings

Old tapes, wax cylinders, war-era broadcasts and family recordings combine noise, distortion, reverb and band limiting. This tool tackles all of them at once to make historic audio listenable.

Pro Premium AI tool — included with any paid plan.

🎧

Drop an audio or video file here

or

MP3, WAV, M4A, FLAC, OGG, AAC

Cleaning your audio…

Before

Tip: press the space bar to toggle Before / After.


How it works

VoiceFixer is a single restoration model trained to handle multiple degradations together — noise, reverberation, clipping and low bandwidth — reconstructing a clean modern-sounding voice from badly damaged source.

What it's good for

  • Family tape and reel transfers
  • Historic speeches and broadcasts
  • Oral-history archives
  • Damaged cassette digitizing

Details

Engine
VoiceFixer
Formats
MP3, WAV, M4A, FLAC, OGG, AAC
Price
Paid plans

Frequently asked questions

Yes — that's exactly what this tool is for. It addresses noise, distortion, reverb and band limiting together rather than making you run four separate passes.

Dramatically for intelligibility, though it's a reconstruction: very degraded source is rebuilt plausibly rather than perfectly. Always keep your original transfer.

For records, run de-click first to remove pops, then this for the overall restoration — they complement each other.

Yes - it is built for early formats whose noise, wow and band limiting stack together. The reconstruction is generative, so faint or buried passages are rebuilt plausibly rather than recovered verbatim; preserve your raw transfer.

It aims to keep the speaker's identity and accent while cleaning the medium's damage. On extremely degraded source the generative stage can smooth period-specific vocal texture slightly, so compare against your transfer.

It is trained on voice, so spoken-word archives restore best. Old music recordings come through, but instrument timbres may be altered since the model expects speech.

Digitize the original at the highest sample rate and bit depth you can and keep that untouched master; feed a high-quality WAV or FLAC in so the restoration has the most detail to work from.

Related tools