How Sequential AI works
By Sequential AI · Updated
Sequential AI is a free AI mastering tool: you upload a track, describe the sound you want in plain English, and an AI assistant processes your file and returns a new version. Your original is never replaced. The result sits beside it, so you can compare the two, keep the new version, or ask for something different.
How does it work, step by step?
- Upload a track. Drop an MP3, WAV, FLAC or OGG file into the Studio.
- Say what you want. Type a request in the chat, in your own words, for example "My track sounds thin, enhance it please."
- The assistant does the work. It measures your audio where it needs to, chooses the right processing tools, and renders a new version of the file.
- Listen and compare. The new version arrives next to your original, so you can switch between the two.
- Keep it or ask again. Accept a version to save it to your track's history, or send another instruction. You can go back to any earlier version.
- Check and share. The release check measures the version you accepted. Then you can share a link, publish it to Discover, or send it to YouTube.
What does it sound like?
Each pair below is one track before and after a single plain-language request. Press play on both and compare.
A thin-sounding funk track
Request: “My track sounds thin, enhance it please.”
Result: Fuller, richer tone
Original
Mastered
A noisy disco track
Request: “Clean up the noise in my track”
Result: Background noise removed
Original
Mastered
A guitar track with mic hum
Request: “My mic has some weird hum, please tackle the noise.”
Result: Mic hum removed
Original
Mastered
What can I ask Sequential AI to do?
Anything about the sound of your track, in everyday language. Behind the chat, the assistant has these kinds of processing to draw on:
| You can ask for | What happens | Example request |
|---|---|---|
| Tone changes | Parts of the frequency range, such as the low end, mids or high-end air, are boosted or cut with an equaliser. | "Make the vocals brighter" |
| Dynamics | Compression, multiband compression and limiting even out or tighten the loud and quiet parts. | "Even out the loud and quiet parts" |
| Warmth | Gentle saturation adds harmonic colour in a tube, tape or transistor style. | "Add warmth to the vocals" |
| Noise removal | Spectral noise reduction takes out hiss, hum and background noise. | "Remove the background noise" |
| Loudness | Overall loudness is set, with a true-peak limiter keeping peaks under a ceiling. | "Make it louder without distorting" |
| Stereo width | The stereo image is widened or narrowed. | "Make the mix feel wider" |
| Reverb | Room or hall ambience is added. | "Add a little room reverb" |
| Pitch and tempo | The key is shifted, or the speed is changed, without affecting the other. | "Slow it down a little" |
| A full master | A complete finishing pass: it corrects problems the track actually has, applies a chosen finish, then sets loudness and a safe peak ceiling. | "Master my track" |
The Studio's One-Click Master offers six finishes: Balanced, Warm, Vintage, Bright, Punchy and Clean. By default a master aims for −14 LUFS integrated loudness with a −1 dBTP true-peak ceiling, the range streaming services commonly recommend.
What is the release check?
After you accept a version, the release check measures it and rates each result Good, Check or Issue. It covers loudness, peak headroom, clipping, dynamics, stereo and mono compatibility, tonal balance, background noise, sample rate and duration. Each rating explains what the number means and how to fix it, and a playback preview shows how streaming platforms will handle your file.
What are its limits?
- It is a beta. Always listen to the result before you keep it. That is why the original stays next to every version.
- Editing cannot restore detail the recording never had, so a low-bitrate MP3 caps everything made from it. Upload your best-quality file.
- The assistant sticks to music, mixing and mastering. It will not answer coding or general questions.
Is it free, and who owns my music?
Sequential AI is currently free. If paid plans are introduced, the pricing and terms will be shown before you are ever charged. You keep ownership of everything you upload, including the processed output, and you are responsible for having the rights to any audio you submit. See the FAQ and the Terms of Service for details.
Who makes Sequential AI?
Sequential AI is built by a team of music-AI researchers and engineers, founded by Prof. Dorien Herremans, who leads the Audio, Music, and AI (AMAAI) Lab at the Singapore University of Technology and Design. Meet them on the team page.