Updated November 12, 2026
To turn a photo into a video with AI, you give an image-to-video model your photo as the first frame and a short prompt describing the motion you want. The model keeps the photo as the opening frame and generates the next few seconds from it. cnaps studio runs five open models that work this way, all on the free plan, so you can try the same photo on several of them and keep the clip that moves the way you pictured.
This guide covers which photos animate well, how to write the motion prompt, which model to pick for which job, and what one real run looked like.
What you need
One photo and one sentence. The photo becomes the first frame, so it sets the framing, lighting, and subject for the whole clip. The sentence tells the model what should move and what should stay still.
Photos that animate well have one clear subject, room around it for camera movement, and lighting that already looks finished. Busy scenes with many small objects, text, or hands close to the camera are harder, because the model has to keep all of them consistent while things move.
How to do it in cnaps studio
- Add your photo. Upload it in a new flow, or generate one first with an image model in the same flow.
- Pick an image-to-video model. Connect the photo to the model's image input. The table below helps you choose.
- Write the motion prompt. Describe the camera move and what moves in the scene, then say what must stay fixed. "Slow push-in, steam rising from the cup, the cup and table stay still" works better than a description of the photo itself.
- Run it and compare. Send the same photo and prompt to two or three models at once and line the clips up in Video Compare.
- Upscale if needed. Add SeedVR2 or FlashVSR after the video model to raise the resolution in the same run.
Which model to use
| Model | Best for | Output |
|---|---|---|
| Wan 2.2 TI2V 5B | Clean product and scene clips | 5 to 6 seconds at 1280x704 or 704x1280 |
| LTX-2.5 | Clips that need sound | 5 seconds at 1920x1088 with 48 kHz stereo audio |
| Cosmos3 Nano | Commercial projects that need sound | Video with synchronized audio, under a license that allows commercial use |
| Helios Base | Longer clips | Up to 60 seconds |
| Helios Distilled | Longer clips, faster | Up to 60 seconds, two to three times faster than Base |
Wan 2.2 TI2V 5B locks your photo as the first frame for the whole generation, which keeps the subject faithful to the original. LTX-2.5 and Cosmos3 Nano generate a soundtrack with the picture. Helios Base and Helios Distilled are the ones to use when five seconds is not enough.
The bottle, cap, and wet stone held their shape across all five seconds, and the light moved across the glass as the camera pushed in slightly. The mist arrived as a spray from the nozzle rather than drifting behind the bottle, which is a good reminder to say where a motion comes from. The whole flow, generating the photo and then the clip, took about three minutes.
Turn a photo into a video in cnaps studio →
Tips for better motion
Describe motion, not the picture. The model already sees the photo. Spend the prompt on what changes: the camera move, light shifting, steam, fabric, water.
Say what must not move. Products, logos, and faces drift less when the prompt says they stay still and keep their shape.
Match the aspect ratio to the destination. Wan 2.2 has a portrait option at 704x1280 for Reels and Shorts. Starting from a photo in the same orientation avoids awkward crops.
Keep the first try short. Five seconds is enough to see whether the motion works before you spend time on longer clips.
Turning many photos into clips
Once a flow works for one photo, it works for the rest. Batch Run, available from the Starter plan, sends a whole folder of photos through the same flow, and the flow can also be called through the API or MCP, for example from Claude. That turns one product clip into a clip for every product in a catalog.
The free plan includes all five models, runs one job at a time, and watermarks output. Paid plans start at $20 a month with Starter, which removes the watermark and adds Batch Run.
Turn your first photo into a video on cnaps studio and compare two models on the same shot.
FAQ
Can I turn a photo into a video for free on cnaps studio?
Yes. All five image-to-video models, Wan 2.2 TI2V 5B, LTX-2.5, Cosmos3 Nano, Helios Base, and Helios Distilled, are on the free plan. Free output carries a watermark and runs one job at a time. Paid plans start at $20 a month with Starter, which removes the watermark and adds Batch Run.
How long does it take to turn a photo into a video?
It depends on the model and the queue. In our test, generating a product photo and turning it into a five-second clip with Wan 2.2 TI2V 5B took about three minutes in one flow, and Helios Distilled runs two to three times faster than Helios Base. Start with a short clip to check the motion before generating longer ones.
Can the video have sound?
Yes. LTX-2.5 generates a 48 kHz stereo soundtrack together with the video, and Cosmos3 Nano also generates synchronized audio. Describe the sound in your prompt, such as quiet rain with no music, so the model matches it to the picture.
Can I make vertical videos for Reels and Shorts?
Yes. Wan 2.2 TI2V 5B has a 704x1280 portrait setting. Start from a vertical photo so the first frame already has the right framing, and the model animates it without cropping.
