Works with any AI video model
Prompts written for Kling, Google Gemini Omni, Sora, Runway, Pika and Grok Imagine.
Upload a video you like and get a ready-to-use AI video prompt in seconds, for Kling, Google Gemini, Sora, Runway or Grok. Free, and your video stays on your device.
Prompts written for Kling, Google Gemini Omni, Sora, Runway, Pika and Grok Imagine.
Only a few still frames are read in your browser. The video itself is never uploaded.
Change any detail, copy the prompt, or turn it into a video in one tap.
5 prompts a day without an account, 15 a day with a free account. No credits used.
An MP4 or MOV from your phone or computer. A Reel, a TikTok or your own clip.
The AI looks at frames from the start, middle and end, and writes one prompt for the whole shot.
Copy it into any AI video tool, or tap Make a video to create it in AI Agent Help.
A video to prompt generator works backwards from a finished clip: it looks at what is on screen and writes the text prompt an AI video model would need to make something similar. It's the fastest way to learn how good prompts are written, and to turn a trend you like into your own original video.
The prompt describes the scene as one continuous shot, because that is what AI video models do best: about 8 to 10 seconds of a single subject, one action, one place and one camera move.
Inside AI Agent Help, tap Make a video and the prompt opens in Text to Video. Choose Grok Imagine (fast), Kling (10 seconds) or Google Gemini Omni (most realistic), and you get a vertical 9:16 video ready to publish or schedule on Instagram. Video is part of the paid plans, and Starter has a 7-day free trial.
Yes. You can use it without an account, up to 5 prompts a day. With a free AI Agent Help account you get 15 a day. It never uses credits.
No. Your browser takes a few still frames from the video and only those are analysed. The video file stays on your device.
Any text-to-video model: Kling, Google Gemini Omni, Sora, Runway, Pika, Luma or Grok Imagine. The prompt is plain English text.
No. AI models create new footage from the prompt, so you get a video in the same style and mood, not a copy. That's what makes it your own.
The prompt follows the language of the page, or your account language when you're signed in. Most AI video models understand English best.
Short clips with one clear scene: cinematic shots, nature, animals, products, 3D characters or aesthetic loops.