AI product launch images and video
Launch assets with an agent are an image and a short video made in one conversation. Your agent picks a hosted model with media__list_models, reads its rate with media__get_model, generates the launch image, passes the image's Drive path into a video model, and hands you both links. A failed generation is never charged.
The job. Pick a hosted model, make a launch image, chain it into a short video, and collect the links.
Paste this into Claude Code, Codex or Cursor
set up https://corespeed.io/SKILL.md (read it raw, e.g. with curl), then pick an image model, make a launch image for our product in our brand colors, turn it into a five-second video with a video model, tell me each model's rate before you run it, and give me both links.One prompt.Three steps.
Set up first. Nothing. No model provider account is needed, and failed, cancelled and expired generations are never charged. To use your logo, upload it to Drive or let the agent upload it.
Ask for the rate first
The agent reads each model's rate before it runs. Video rates differ most.
Chain the image into the video
The image's Drive path is the video model's input. It never leaves CoreSpeed between steps.
Keep what you like
Signed links last 12 hours. Move the files out of tmp/ to keep them.
The tools it calls
media__list_models· Find an image model and a video modelmedia__get_model· Read each rate before runningmedia__generate· Make the image, then the videomedia__get_result· Collect the video and fresh linksmedia__list_jobs· Find a job if the client timed outdrive__move· Keep the files out of tmp/
Every tool and its price: Image generation, Video generation