As of 13 August 2026, AI can create a backing track.
This still needs a person who signs their name to it.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill neededchat-fluent
Who has to check ityou
What the alternative costsA human musician or recording studio is an alternative, but the supplied tool data gives no price for it.
If this goes wrong: the groove, key or arrangement does not suit your song, so you regenerate it or edit the parts before using it.
What to actually do
Hand it to a person
The route this page recommends
A person who owns the outcome does this end to end, worth it when the failure is dear.
Do it yourself
Second choiceA chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.
How to actually do it
- Write down the song's key, tempo, time signature, section order, approximate bar counts, mood and the instruments you want or want to avoid.
- Confirm that the song and any recording you plan to use are yours or that you have permission to use them, then open SOUNDRAW.
- Paste the prompt and replace every bracketed slot with your song details, including the performance use.
- Generate several instrumental versions and choose the one whose key, tempo, groove and section changes fit your singing or playing.
- Export the chosen track and open it beside your vocal or instrument recording in a DAW, or play both together from separate devices.
- Check the intro count-in, key, tempo, chord movement, section lengths, vocal space, transitions and ending, then regenerate or edit the track where one of these does not fit.
- Keep the final audio and the tool's usage or licence information with your project before sharing or performing it.
Prompt
Create an instrumental backing track for an original song or a song I have permission to use. Do not include vocals, a lead melody that competes with the singer, samples with unclear rights or a direct imitation of a named artist. Use these specifications: Genre and mood: [genre and mood] Key: [key] Tempo: [BPM or approximate pace] Time signature: [time signature] Length: [length] Song sections and bar counts: [intro, verse, pre-chorus, chorus, bridge, outro and any repeats] Main instruments: [instruments] Instruments to avoid: [instruments] Energy and dynamics: [quiet or sparse sections, build-ups and strongest section] Performance use: [singing, live performance, rehearsal, video or other use] Keep the harmony and rhythm clear enough for a singer to follow. Leave space in the mid-range for vocals, make transitions obvious, and provide an ending rather than cutting off. If you cannot render audio, return a complete arrangement plan with chord progression, bar counts, instrument entries, drum pattern, bass movement and exact tempo for use in an AI music tool or DAW.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
Use a tool built for this
The distant third
What it gets wrong
- It does not know the emotional detail you hear in your head unless you express it through specific musical instructions and examples.
- It may produce a plausible arrangement with weak chord movement, repetitive drums or transitions that do not support your actual melody.
- It cannot reliably judge whether the track sits in the right register for your voice or instrument without your recording and musical checks.
- It does not clear copyright in a melody, recording or reference track for you.
- It does not replace a musician who can respond to live timing, phrasing and last-minute changes during a performance.
Even on a YES, the friction has a name: taste, context depth and verification cost.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 2 |
| Verification | 1 |
| Liability | 2 |
| Effort delta | 2 |
| Total | 9 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can AI make an instrumental backing track?
- Yes. Tools such as SOUNDRAW and Mubert can generate instrumental music from details such as mood, length and intended use, and SOUNDRAW provides editable tracks. You still need to specify the key, tempo, sections and space needed for your vocals.
- Can AI make a backing track for my song?
- Yes, if you provide enough musical information and use material you own or have permission to use. Give it the key, tempo, chord progression, section order and instruments, then check the result against your actual melody rather than trusting the description alone.
- Can AI remove the vocals and leave a backing track?
- A stem-splitting tool such as LALAL.AI can separate vocals and instruments from a recording, but the result is not always clean. It is different from creating a new backing track, and you need permission to process and use the recording.
- Can I use an AI backing track in public?
- That depends on the tool's licence and on whether the underlying song or recording is yours or cleared for use. Check the current usage terms for the specific track and keep the licence information with your project before publishing or performing it.
Nearby answers
- Can AI help me build a home recording studio in the UK?YES
- Can AI change the tempo of a song without changing its pitch?YES
- Can AI check whether music is copyrighted?NO
- Can AI help me choose a microphone for my podcast?YES
- Can AI help me choose a microphone for singing?YES
- Can AI choose background music for my video?YES
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.