make a video from a prompt · 666 AI
One key, 56 video models, all at your fingertips
Veo needs Google Cloud, Kling requires mainland China ID verification, Seedance needs another account — just getting set up can eat up a week of your time. We've already handled all of that for you.
Start free
Three steps
- Sign up for one account and get one API Key
- Plug the endpoint and the key into your code — OpenAI-compatible syntax
- Switch models by changing just the model name, nothing else
Say something like thisUse kling-video/v3/pro/image-to-video to generate a 5-second video from the image I uploaded, and have the person in the frame turn to look at the camera.
After Sora shut down, migration shouldn't take another week
Upstream providers may go offline tomorrow, change APIs the day after, or hike prices next week — we switch routes on our end, and you don't change a single line of code. What you rely on is one key, not one vendor's account.
56 models, zero application processes
28 from Veo (Google), 12 from Kling (Kuaishou), 5 from Seedance (ByteDance), plus Wan, MiniMax, Grok, and PixVerse. All under the same key, same syntax.
Zero setupNo installs, no code. If you can type, you're set.
Top models, freeFull GPT, Claude and Gemini with a daily quota.
Never pick again250+ models; the right one answers.
Starts at a centPer call. Idle days cost nothing.
Models and live prices
Questions
Is it OpenAI-compatible?Yes. Whatever way you call GPT today, that's how you call video models now.
Do I need an overseas account or a foreign card?No. Sign up and you're ready to go.
How is billing calculated?By video duration in seconds. No charge for failed requests.
Try it now, for nothing
Sign up and the free quota is there. No card, nothing to install.
Start free