Voiceover Studio

Voiceover Studio is a self-hosted web app that turns a raw narration recording into a finished documentary-style mix.

You give it the voice and, if you want, a music track. It fixes the pacing of the pauses, keeps the music playing quietly under the speech, raises the music during meaningful silences, and exports a mastered MP3 or WAV. Your recordings stay on your own machine.

How it works

1. Upload the voiceover (MP3, WAV, M4A, FLAC or OGG; up to 4 GB and 6 hours).
2. Choose music: upload a track, reuse a saved one, generate one with AI, or pick No music.
3. Pick how loud the music is while you speak: Soft, Slightly audible or More audible. Two sliders let you fine-tune the music level under speech and in pauses.
4. Click Create documentary mix, compare Polished against Original, and download.

Jobs keep running in the background, so refreshing the browser doesn’t stop them. Projects save automatically.

Documentary pacing

• Finds the natural breaks in the narration and gives the important ones room to breathe, the way a documentary editor would.
• Leaves uncertain spots, lists and numbers exactly as recorded.
• Works in English, Urdu and Hindi.

Mixing and mastering

• Keeps a steady music bed under speech and lifts it smoothly in the longer pauses, without pumping on breaths.
• Lines up pauses with the music so the voice comes back in on the beat.
• Loops the music, adds fades, cleans up the voice and masters the result to a consistent, broadcast-ready loudness.

AI sound designer (optional)

• AI pause editor: reads the story, finds its key moments (hook, reveal, time or place jump, close) and places dramatic holds there. If it can’t, the automatic pauses are used and a note explains why.
• “Pauses in this mix” report: lists every changed pause with the words on either side, its old and new length, and why it changed. Click a row to hear it.
• AI music: composes an original instrumental track sized to your edited voiceover.
• Music styles: pick a direction for the score, such as a tense investigation or a classic documentary feel.
• Long-form mode for recordings over 10 minutes: splits the story into chapters and scores it with a set of recurring themes that crossfade between sections.

Music from a prompt

• A separate Music page for making tracks directly from a description.
• Settings: duration (5–600 s), BPM, key, time signature, 1–4 variants, and optional lyrics with a vocal language.
• Regenerate the same take or a variation of it, or press Surprise me.
• Results go to your personal music library, and you can put any of them into a project.

Render on Server or on Device

• Server: the normal render on the studio machine.
• Device: the final render runs in your browser instead. Both give the same result.

Advanced editor

• Edit or lock individual pauses, supply a script, and change the language.
• Waveform view, undo/redo, and a mode for recordings that already have voice and music combined (pause edits only).
• Export a project and summary as a ZIP.