Studio notes
How to make an AI rap video from a photo
Prepare one portrait and a short topic, choose a rap language and video shape, and understand the preview limits before creating a personalized clip.
· Oct 23, 2026 · Last updated: Oct 26, 2026
Start with a photo and one clear idea
The original-rap workflow is designed around one adult portrait and a short topic, such as a birthday, a promotion or a friendly joke. The goal is an approximately 15-second solo rap performance with newly generated audio inspired by that topic. You do not need to upload a song or write a finished verse.
Current availability: the studio is in preview. You can prepare your photo and topic locally and watch reviewed samples. Generation and payment remain closed until the relevant workflow, output shape and quality are approved. The instructions below distinguish what you can prepare now from what a future generation would do.
1. Choose a clear adult portrait
Use one adult whose photo you are allowed to use. Prefer a sharp face, even light and visible eyes and mouth. Start with a JPG, PNG or WebP file up to 8 MB. Choosing a photo creates a local browser preview; it does not submit a model job or start a charge.
Read the photo checklist before changing lots of settings. A good input helps, but does not guarantee exact identity through every frame.
2. Describe a topic, not an exact script
For a birthday, mention the occasion and one personal detail. For a milestone, name the achievement and the feeling you want to celebrate. For a friendly roast, choose an everyday habit the person is happy to joke about.
For example: “A cheerful birthday rap for an adult friend who always brings pizza to game night.” This is a topic brief, not a transcript or a promise of the words the video will say. Names, rhyme, pronunciation and wording can change.
Choose a vocal language separately from the site language. English, German, French, Spanish and Mandarin Chinese are selectable directions; availability of a selector does not establish output quality in every language. We do not offer voice cloning, uploaded songs or exact-word lyrics.
3. Choose the output shape and check the mode
Use 9:16 for a vertical greeting or 16:9 for a wider screen. Each shape needs separate validation; a reframed portrait preview is not proof of a native landscape generation.
Original rap uses a topic to request new audio. A fixed-audio template uses an existing performance soundtrack. The Orange Studio sample belongs to the latter test workflow: it shows portrait replacement with fixed reference audio. It is not evidence of a newly personalized birthday song.
4. Check availability and price before starting
There is no free generation tier. The pricing page explains paid video packs and plans; the applicable amount must be shown before a job starts. Internal model cost estimates are not customer prices, and a standard template allowance is not automatically an approved price for original rap.
Once a workflow is open, sign in and review the current offer. A slow response is not permission to submit the same job repeatedly. Check My videos for progress, and distinguish a confirmed failed job from a submission with an uncertain status.
5. Review the whole clip with sound
Check identity, hands, mouth movement, audio, the topic and the ending. Generated video can contain artifacts and cannot guarantee word-perfect lip-sync. Save your MP4 while it is available. Do not describe an AI performance as a real recording of the person.
Read the real model test for known defects. For a concrete idea, try the birthday topic guide or friendly roast guide. Questions about a job belong in support, without posting private media links publicly.