OpenModels
AI video generator — prompt or still to video, long video to shorts
Generate video from a prompt or a single photo on the models running on our own GPUs, or on Kling and Seedance with a plan. Feed in a long video and OpenModels finds the moments worth clipping and cuts them into captioned shorts. Free accounts start with daily GPU credits and no card.
Which plan: 20 free credits every day — GPU credits for the models on our own GPUs, growing to 40 a day when you come back. No credit card required. On video that pays for the LTX and MiniMax lanes, billed per five-second block, with the price of the clip on the button before you press it. Shorts extraction, editing and the premium video models come with a plan.
Generate
- Text to video and image to video on LTX and MiniMax (GPU credits) or Kling and Seedance (plan).
- Motion control: pick the camera move — orbit, push-in, pan — and keep the same subject.
- Character and reference clips: hand the model a face or a short clip and it stays consistent across shots.
- Upscale to 720p or 1080p and smooth to 60 fps.
Analyze a long video
- Describe the moment you want in words — the duel, the reveal, where the music swells — and get the timestamps back.
- Ask for the trending moments and get a ranked list of short-form candidates.
- Extract the clips with captions burned in, ready for Shorts, Reels and TikTok.
- Pays per sampled frame in GPU credits, quoted against your own file before anything runs, plus 1 credit per moment ask; the export itself is free, and anything left undescribed is refunded.
How it works
- Describe it Type a prompt. Attach a photo if you want to edit it, restage it or turn it into video.
- Pick the models Tick one or several. Models on our own GPUs spend the daily GPU credits; premium models spend Premium model credits, and the price is shown on the button before you press it.
- Generate Every result lands in one album, side by side, so you can compare models on the same prompt and keep the one you like.
Generate a video
Video models
- Seedance 2.0 Mini — Audio is generated with the frames, so footsteps and room tone match what is on screen.
- Kling 3.0 Pro — Weight, balance and follow-through survive — the failure mode most video models still have.
- Google Veo 3.1 — Dialogue, ambience and effects are generated with the frames rather than dubbed over them.
- Grok 2 Imagine — One of the lower-cost ways here to turn a picture you already generated into a moving shot.
- All models