Can AI make a song from a voice memo or a video?

Yes in Suno, no in Sume. Suno's v6 post says you can start a song from text, audio, images or video, including a voice memo. Sume's music API takes a text prompt and one optional image; it has no field for an audio file or a video, so a voice memo has to be described in words instead.

Suno's wording is quoted from Introducing v6. Sume's fields come from the Music 1.0 and Music Router docs. Both were read 2026-09-29.

What does Suno say it can start from?

The post is dated Sep 9, 2026, and lists the inputs in one sentence with an example request.

What can Sume's music API take as input?

The docs give two inputs: a prompt of 1 to 5000 characters and an optional image_url, which must be a public HTTPS image. Nothing in either docs page lists an audio or video input. The docs call the image optional visual conditioning.

How do I use a voice memo anyway?

Listen to it and write down what you hear as prompt text: the tempo, the key if you know it, the mood, the instruments and the words you sang. The Music 1.0 docs say the prompt steers the song, so the more of the memo you put into words, the closer the new track can get. It will be a new song that follows your description, not your recording turned into music.

• Write the lyrics you sang into the prompt under section labels such as [Verse] and [Chorus].
• Say the length in the prompt. duration and duration_seconds are rejected.
• Read result.lyrics after the job, which the docs call model-reported, and listen before you rely on it.

Does the image field replace the audio input?

No. image_url is separate from anything you sing or hum, and the docs describe no way to send a recording next to it. If you want your own recorded voice in the final piece, you can lay it over a generated track afterwards; mixing a voice with background music shows how the timeline handles that.

How do I use a video?

Pick one still from the video, put it in image_url, and describe the pacing in the prompt. The music for a video post covers that route, and image to music covers the image field on its own.