How to make a Tiny Desk AI video
A practical photo-to-performance guide for your first 19-second session.
1. Start with the right portrait
Choose a sharp, front-facing photo with one person in even lighting. The face should be clearly visible. Avoid group shots, heavy filters, face-covering accessories and tiny screenshots. Use only a photo you have permission to process.
2. Set up your session
Open the studio and select a JPG, PNG or WebP image up to 10 MB. You can replace or remove your photo before submitting. A local preview stays in your browser until you start generation.
3. Review your session and check the cost
Each 19-second session uses 70 credits with a 480p target output. There is no quality setting to choose. Identity consistency and visual quality depend on the source image and the AI result. Purchased credits never expire.
4. Generate the full performance
When the studio is available, sign in, confirm permission and submit once. Your portrait replaces the lead vocalist while the reference controls band positions, camera changes and performance timing. The prompt requests a fresh fictional adult supporting cast, keeping appearances consistent within the clip; results vary. The matching configured soundtrack is attached after the visual edit finishes.
5. Review, download and share
Keep the page open or return through My videos. A completed video appears only after final processing. Play it before downloading the MP4 and check the face in close-ups, profiles and shot changes. If generation fails, the reserved credits are returned to your balance.
What if the studio is not open yet?
You can try photo selection and preview before buying credits. Sign in with Google and purchase credits to submit a generation. The homepage demo is a supplied example; individual results may vary.