Generating video, images, voice & music
Ask the AI in plain words, or open the 🎛 panel to pick exact models — check several to compare outputs side by side. Billed per delivered result.
Two ways to generate media, same engine underneath:
Just ask
Type it anywhere the AI listens — the chat bar or the ✨ tool: "create a 5-second clip of waves at dusk", "generate a voiceover saying …". Name an engine and it's honored ("use Kling"); ask "which video models can I use?" for the menu with prices.
Or pick models yourself (🎛)
Click the ✨ tool on empty canvas, then 🎛 models… in the prompt box. Choose what you're making (video / image / voiceover / music), checkbox the models you want — prices in ⚡ next to each — write one prompt, and hit Generate. Checking several models is the compare feature: each renders its own card, labeled with its model, so you can play them side by side and drag the winner onto your timeline.
Pricing & reliability
Generation is billed per delivered result — the price shown is charged only when your file actually arrives; failures and timeouts cost nothing (see credits). Renders run server-side, so closing the tab is fine — cards update when results land, showing the first frame of finished videos.
Generated assets behave like any media: drag onto a timeline lane as b-roll, or ask the agent to place them (see edits & segments).