How to Make an AI Rap Video From Two Photos (Step by Step)

An AI rap video used to mean a film crew, a studio session and a lot of favours. Today you can get a surprisingly convincing result from two photos and a few sentences of direction. This guide walks through the whole process on RapDuo, from picking images to sharing the finished clip, and explains the small decisions that separate a forgettable output from one people actually replay.

If you would rather skip straight to trying it, the generator on our homepage lets you sign up free and get 3 credits, enough for your first video. Keep this page open as a checklist.

What you are actually making

A RapDuo video is a short clip in which the people in your photos perform an original rap. Three things are generated for you: the lyrics (or your own words, if you write them), the vocal and beat, and the video itself, with faces animated and lip-synced to the performance. Nothing is copied from a real song or a real artist. The music, voices and visuals are all new, which is why the results are safe to post as AI-generated creative work.

The typical clip runs between 5 and 30 seconds depending on the tier and settings you choose. Short is a feature. A tight 12-second verse with a strong hook is more shareable than a rambling minute.

Step 1: Choose two photos

Photos are the single biggest factor in quality. The model has to invent motion, expressions and mouth shapes from a still image, so the clearer the input, the more believable the output. A good rule: pick a photo where you could easily describe what the person is feeling.

• Face fully visible. Front-on or a gentle three-quarter angle. Avoid sunglasses, hands over the mouth, masks and heavy shadows across the eyes.
• Sharp and evenly lit. Daylight near a window beats a flash-lit party photo. Motion blur becomes strange warping later.
• One person per photo. Two photos, two performers. Group shots confuse the animation.
• Decent resolution. A phone photo is plenty. Screenshots of screenshots are not.
• Adults who agreed to it. Only use photos of yourself or of adults who are happy to appear. Never use photos of minors.

We go deeper on this in our guide to the best photos for an AI rap video, including what to do with photos that are not perfect.

Step 2: Pick a stage

The stage is the world your performers rap in. RapDuo includes four looks, each designed to feel like a real place rather than a green-screen backdrop.

• Hotel Lobby: the warm, orange-lit lobby booth with a single fixed camera. This is the look behind the viral trend. Read how to do the hotel lobby trend for details.
• Luxury Lobby: marble, gold trim and a chandelier. It works well for boastful, over-the-top lyrics.
• Studio Booth: a dim vocal booth with neon and pop-filter microphones. The natural choice for anything that should feel like a recording session.
• Street Cypher: a night block, streetlights and a circle of friends. Gritty, film-grain energy.

Choose the stage that matches the joke or the mood, not the one that looks coolest in isolation. A birthday shout-out in a moody street cypher can be great comedy, but only if you mean it.

Step 3: Decide the words

You have two options. You can describe the topic and let the AI write the lyrics, or you can write the lines yourself. Both work; they produce different kinds of videos.

Let the AI write it

Give it a topic, a tone and two or three concrete details. "A roast of my brother Sam who is always late, playful, mentions his red scooter" gives the writer something real to work with. Specifics create punchlines. Vague requests such as "a funny rap" produce generic rhymes.

Write it yourself

If you want exact lines, keep them short. Around 8 to 12 syllables per line rhymes and lip-syncs best, and four to six lines is enough for a clip. Read each line aloud once. If you run out of breath, the performer will too. Avoid tongue-twisters and long strings of names, which are the most common source of odd mouth movements.

Need ideas? 30 creative rap video prompts is a good place to start.

Step 4: Choose language, tier and quality

RapDuo can perform in English, Chinese, Spanish, Portuguese, French and Italian. Pick the language your lyrics are written in and keep it consistent; mixing languages inside one line often trips up pronunciation.

There are four quality tiers. Lite is the fastest and most affordable, ideal for drafts, memes and social posts. Pro adds sharper motion and more faithful faces. Ultra raises the detail and adds a true 1080p option, and Ultra Plus is the flagship for when the video is the main event, such as a birthday gift or a wedding toast. You can see the credit cost before you confirm anything, and our credits and pricing explainer breaks down exactly what each option costs.

Step 5: Generate and check the scene preview

Sign up free and verify your email to get 3 credits, exactly one 5-second Lite 480p video. One click runs two steps: your two photos are composed onto the stage as a scene preview, then that scene becomes the first frame of the lip-synced video. Look for three things.

1. Do the faces still look like the people in your photos?
2. Does the mouth movement roughly follow the words?
3. Does the beat and mood match what you had in mind?

If any answer is no, change one thing at a time. Swap a photo, shorten the lyrics, or choose a different stage. Changing everything at once makes it impossible to learn what helped.

Step 6: Download and share

When the video is ready, download the MP4. It has no watermark and is stored for 30 days. If a render fails or is blocked by our safety check, the credits are returned automatically. To make another version, use credits from a pack or plan, so check the scene preview carefully.

Common mistakes and quick fixes

• The face looks rubbery. The photo is too dark or too angled. Try a brighter, more frontal image.
• Lip-sync is off on fast lines. Simplify the lyrics and cut the syllable count.
• Everyone looks the same. Use two distinct photos with different framing, and describe each performer in the prompt.
• The joke does not land. Add one specific detail that only the recipient would recognise. Personal beats polished.
• The clip is too long. Trim to a single verse. Ending early is better than dragging.

Sharing it responsibly

Label your video as AI-generated when you post it, and only share clips of people who have agreed to appear. It keeps the fun fun. Our Acceptable Use Policy covers the rules in plain language.

Ready to try it?

Upload two photos, pick a stage, and get your first video in a few minutes. Start on the RapDuo homepage. Signing up is free and needs no card.