THE SONGWRITER’S NOTEBOOK
Make your next song with an AI Music Generator
Start with a scene, a lyric or a sound. The AI Music Generator brings your musical direction, lyrics and audio settings together, so you can shape an idea and listen to where it leads.
01 / FIND THE SOUND
Describe what you want to hear.
A style, a few instruments and a mood give the song a starting point. Describe the feeling in one or two clear sentences.
An intimate folk song with a warm female lead, fingerpicked acoustic guitar and brushed drums. Keep the verses restrained, then lift the final chorus with soft harmonies.
How to write a clearer music prompt
Describe the style, emotional setting and main instruments in plain language. Give the AI Music Generator a sense of movement too: perhaps the verse stays close and quiet before the chorus opens up. Specific contrasts within a song are useful. A long list of unrelated adjectives is harder to assess when you listen back. Keep the most important part of the idea near the beginning of the description.
Start with a small idea you can judge by ear. A quiet piano theme for a film opening needs a different brief from an upbeat chorus for a short video. Pick the version that serves the project, then change one part of the brief on the next attempt. This gives you a clearer basis for comparing takes than adding every possible genre, instrument and mood at once.
That is an example brief, not a promise about a finished recording. Use it as a starting point and replace the details that do not fit your song. The genre, mood and instrument fields add context to the description; they do not require specialist terminology. For more about the model behind this workspace, read the MiniMax Music 3.0 overview.
Explore MiniMax Music 3.0
02 / PUT IT INTO WORDS
A few words. A first verse.
Bring your own lyrics, draft them with the writing assistant, or leave room for the model to write. Choose Instrumental when the music should speak for itself.
Working with lyrics and song sections
The AI Music Generator can work with lyrics you already have. Separate verses and choruses with section labels so the intended shape is readable. Shorter lines and a chorus with a clear central phrase give you something concrete to evaluate after generation. The lyrics editor keeps long drafts inside a scrolling area, so the next action stays within reach while you revise a verse.
If you are starting from a topic, use the lyrics assistant to draft words, then read them before generating audio. Rewriting an existing draft asks for confirmation because it replaces the text. Arrange sections keeps the words and adjusts their organization. You can also leave lyrics blank and let the model write them. Switch to Instrumental when the piece should have no sung words; your draft stays available when you return to Song.
03 / SHAPE THE PERFORMANCE
Give the sound a direction.
Choose one voice, then add the instruments and mood you hear in your head. BPM and key guide the prompt; they do not guarantee an exact musical match.
What each music setting actually changes
Voice, tempo, key and structure help the AI Music Generator interpret the brief. Choose one voice direction: Auto, Male, Female or Duet. A duet is one creative choice for two vocal parts, not two separate toggles. The content language guides the vocal request and writing assistant; it does not change the language of the website. Listen to pronunciation and phrasing before choosing a take.
BPM and key are written into the prompt as preferences. They are not dedicated MiniMax API fields, and the result may differ from the requested tempo or tonality. If the recording must match another track exactly, check it in your audio editor before combining them. Audio format, sample rate and bitrate are different: the page sends these as output settings supported by the provider.
| Setting | What it changes |
|---|---|
| Voice, BPM and key | Creative guidance in the prompt. Check the generated performance by listening. |
| Format and sample rate | Native audio output settings. Choose a file type that fits the next tool in your workflow. |
| Song structure and lyrics template | Arrangement guidance and text organization. They do not define exact section timestamps. |
The right file for what comes next.
mp3
An easy choice for everyday listening, browser previews and sharing.
wav
Take the recording into an audio editor that accepts WAV files.
pcm
Raw audio for compatible tools. It may not play directly in a browser.
Compare formats and output settings
The AI Music Generator offers MP3, WAV and PCM because all three appear in the provider’s audio settings. MP3 is convenient for everyday listening and sharing. WAV is useful when the next step is an audio editor that accepts it. PCM is raw audio and may need a compatible player or editor; browser playback support is separate from whether the provider can return that format.
The default selection is 44.1 kHz, 256 kbps and MP3. Choose other settings for a specific delivery requirement rather than expecting a format switch to improve the composition. A higher bitrate does not decide the melody or fix a lyric. MiniMax’s official API reference documents the available audio fields, including the three formats, so you can check the provider contract directly.
Your song can become a visual story.
Listen to the finished track, then give it a setting, a mood and a sequence of scenes. Start an MV when the music is ready.
Turn a finished song into an MVFrom listening to planning an MV
After the AI Music Generator saves a successful recording, listen in the results panel or open your music library. Check the introduction, the transition into the chorus and the ending. A strong first few seconds may suit a short clip, while a slower build can need more room. Keep the original brief nearby so the next generation responds to a specific observation rather than a vague request to make everything better.
When a track is ready for a visual treatment, choose it in the video generator. That page adds visual style, frame and length choices, then asks you to confirm before rendering scenes. The professional studio provides more control over the storyboard and individual shots. Music creation and video creation remain separate decisions, so you can finish the sound before taking on the visual work.
AI Music Generator questions, answered
Can I use the AI Music Generator without writing lyrics?
Yes. Leave the lyrics field empty for the model to write words from the musical direction, or select Instrumental for a piece without vocals. If you want to review the words before audio generation, use the writing assistant first. These are separate ways to start; choose the one that matches how much control you want over the text.
Does the AI Music Generator follow BPM and key exactly?
No exact match is promised. Tempo and key selections become language in the prompt, while the model produces the performance. Treat them as a direction and check the recording before using it in a project with strict musical requirements. Leaving either setting on Auto avoids adding that preference to the request.
Which audio formats can I choose?
MP3, WAV and PCM are available in the output controls. The request sends the chosen format to MiniMax rather than converting every selection to MP3. PCM files may not play directly in a browser. Choose MP3 for convenient previewing, or check that your editing software accepts the other format you need.
Do I need an account and credits to generate a song?
You can explore the form before signing in. Generating audio and using the writing assistant require an account. The music page displays 50 credits per successful generation; check your balance and the pricing page before starting. The ability to open the AI Music Generator does not mean audio generation is free or unlimited.
Will changing the voice or mode erase my lyrics?
Changing the voice keeps the current lyrics. Switching to Instrumental hides the lyric controls and leaves the draft in this open page; switching back to Song restores it. An AI rewrite is different and asks before replacing existing words. Copy any draft you need to keep beyond the current page session before leaving or refreshing.