STEP-BY-STEP TUTORIAL & PROMPT PLAYBOOK
To make an AI rap video from photos in RapPrism, upload 1 portrait photo (Solo Mode) or 2 separate portrait photos (AI Rap Duo Mode), write a 1–2 sentence rap topic with names and inside jokes, pick your Flow Energy and Scene, and click Generate. In 1 to 3 minutes, you get a 9:16 vertical 720p MP4 with original English rap vocals and synced lip movement.
01 / 3-STEP PRODUCTION WORKFLOW
Follow these three concrete steps to go from raw camera roll portraits to a downloadable 9:16 rap performance.
Select AI Rap Duo if two people are trading bars, or Solo Mode for a single performer. Upload one crisp, front-facing portrait per slot where the eyes, nose, and closed or neutral mouth are clearly visible without hands, sunglasses, or heavy shadows.
Instead of typing a single generic word like 'birthday', write who is in the clip, what you are celebrating or roasting, and two specific details (e.g., 'Sam turning 28, always late to brunch, burns toast'). Then select Hype, Playful, or Smooth energy and a Studio, Street, or Stage backdrop.
Check the portrait authorization box and submit your job (50 Credits). RapPrism synthesizes custom English rap bars, vocal cadence, and lip-synced motion. Preview the 9:16 vertical clip in your browser and download the watermark-free 720p MP4.
02 / INPUT QUALITY CHECKLIST
Lip-sync accuracy and facial identity preservation depend heavily on your source portrait. Use this checklist before uploading.
03 / PROMPT FORMULA & COPYABLE SCRIPTS
Great AI rap verses come from specific details. Combine these four building blocks inside the 240-character topic box:
[1. Performers & Relationship] + [2. Occasion / Battle Topic] + [3. Two Specific Habits or Inside Jokes] + [4. Closing Vibe]
“Make a funny birthday rap video for my friend.”
“Two best friends roasting Leo on his 30th birthday for buying espresso gear he never uses and still showing up 20 minutes late.”
“Two best friends trading funny birthday roast bars for Sam turning 28, hyping up the weekend party while joking about his terrible cooking.”
“Duo rap battle between two roommates arguing over who actually cleans the apartment vs who just orders late-night takeout every day.”
“Couple celebrating 3 years together, rapping about surviving missed flights in Tokyo, midnight ramen runs, and building their dream life.”
“Two co-founders celebrating shipping version 2.0 after 60 late nights of debugging, high-energy stage performance shouting out the team.”
“Solo creator introducing a brand-new AI SaaS launch from a warm neon recording booth, confident smooth boom-bap cadence.”
“Solo victory rap flexing an undefeated fantasy football championship trophy and calling out the group chat runners-up.”
04 / TROUBLESHOOTING & FIXES
If a render looks off or fails to complete, here is what caused it and how to fix it on your next run.
Cause: The source photo had an open laughing mouth, heavy shadow under the nose, or a side-angle jaw. Fix: Replace with a well-lit, front-facing portrait where the lips are relaxed and closed.
Cause: One of the uploaded files contained multiple people in the background. Fix: Crop each photo tightly around a single person's head and shoulders before uploading to Slot 1 and Slot 2.
Cause: The topic prompt was under 5 words. Fix: Include at least 1 first name, the specific event, and 2 concrete habits or jokes (80–180 characters is the sweet spot).
Cause: Temporary upstream GPU queue congestion or safety filter rejection. Fix: Check that your prompt and portraits follow content guidelines; your 50 Credits are automatically refunded to your balance so you can retry immediately.
05 / TUTORIAL FAQ
No. RapPrism is purpose-built to generate both original English rap vocals and lip-synced 9:16 video directly from 1–2 portrait photos and a text topic. It does not accept external MP3 song uploads.
No. Always upload two separate single-person portrait photos — Rapper 01 in Slot 1 and Rapper 02 in Slot 2. Uploading a single group photo reduces facial tracking accuracy.
Generated rap vocals and rhymes are performed in English for optimal rhythm and phoneme lip-sync. You can write your topic prompt in English or Chinese.
Never. While 50 Credits are reserved when a task starts, our task service automatically revokes the charge and refunds all 50 Credits if the upstream render fails.