GET /v1/images1 credit/callList every image a page loads - img, srcset, picture, CSS backgrounds, favicons, og:image - one API call, tagged by source.
One authenticated GET request against any live URL - clean, typed JSON back.
GET /v1/imagesconst res = await fetch(
`https://miromiro.app/api/v1/images?url=stripe.com&access_key=${process.env.MIROMIRO_API_KEY}`
)
const data = await res.json()Catches CSS background-image, favicons, and og:image - sources a plain <img> scrape silently drops.
Every endpoint plugs straight into your coding agent - as a native MCP tool or a plain-English prompt. No SDK, no glue code.
Your agent calls it directly, mid-task, and gets typed JSON back.
// Cursor · Claude Code · Windsurf
extract_assets({ url: "stripe.com" })
No MCP? Drop this into any agent - it wires up the API itself.
“List every image on stripe.com and download the product shots.”
The MiroMiro Image Extraction API lists every image a page loads - img src and srcset, <picture> sources, CSS background-image, favicons and apple-touch-icons, and og:image / twitter:image - as resolved URLs, each tagged by the source it came from. One GET request, complete inventory.
One GET request with any live URL - e.g. stripe.com.
Clean, typed JSON back - resolved, not raw HTML to parse.
MCP or a prompt wires it into your build - no glue code.
Pull every image from a legacy site in one call for a migration.
Assemble product imagery and brand assets programmatically.
Feed images into downstream processing or moderation.
img src and srcset, <picture> sources, CSS background-image, favicons/apple-touch-icons, and og:image / twitter:image - each tagged by where it was found.
Image extraction is 1 credit per call. Cached and failed calls are free.
100 free credits every month, no credit card. One key unlocks every endpoint - brand, tokens, images, SVGs, fonts, and Lottie. Upgrade only when you outgrow the free tier.