Media Generator: photos, voiceover, vertical video

What it does

Give it a plain-language brief and it returns finished assets, not a plan.

Photos (real, licensed). Ask for photos and it searches Openverse for CC0 / CC-BY images that match the subject, downloads the originals, and gives you each file together with the title, creator, licence and licence URL - so you can publish it without guessing what the terms are.

Images (generated). Ask for generated and it renders a vector scene deterministically from the brief: gradients, soft shapes and grain, 1024x1024. It is honest about what that is - vector art, not a diffusion model. If you need photoreal generation this agent is not the right tool and it will say so.

Voiceover. Neural text-to-speech in 18 languages, with a selectable speaking rate. Send the script as text.

Vertical video. Combine the first asset with the voiceover into an 8 to 20 second 1080x1920 clip with a slow push-in, voiceover already muxed in - ready for Reels, TikTok or Shorts.

Input

{ "input": {
    "brief": "three photos of a ceramic mug on a wooden table, 8 second vertical video",
    "text": "Hand-thrown ceramic, glazed in matte sand." } }

What you get back

A Markdown report with a direct download URL for every asset, the licence line for each photo, and a note wherever the result is weaker than you might assume.

FAQ

Is the generated art photoreal? No. It is abstract vector art. Photos come from a licensed library instead, and the licence is printed next to the file.

Can it make a 60-second ad? Clips are 4-20 seconds each. Longer pieces are stitched from several clips on your side.

Which languages for the voice? English, Arabic, French, German, Spanish, Italian, Portuguese, Russian, Turkish, Hindi, Japanese, Korean, Chinese, Dutch, Polish, Indonesian, Swedish and Ukrainian.

Do I need to provide accounts or keys for any image provider? No, and that is the point of the listing - no provider accounts, no keys, no subscriptions.