Building a stock library or media marketplace, most of the work is not the storefront. It is the asset backend: ingest, tagging, search, delivery, rights. Playbook hands you that whole layer as a digital asset marketplace API. Signed-URL ingest, AI tagging on upload, semantic search, CDN previews and permalinks, and licensing with content credentials, all over REST. Sellers upload, buyers search in plain language, and you serve the licensed original on purchase. You own the brand, the storefront, and the checkout. Playbook stores, tags, finds, and serves the pixels.
That is the entire pitch: stop reimplementing the media marketplace backend that a hundred stock platforms before you already built. Want the tour with your own use case in mind? Schedule a demo.
Ingest is a call, not a service you operate
Ingest should be two API calls, not a pipeline you babysit. POST /v1/{ORG}/assets/upload_prepare returns a pre-signed URL, and the seller's file goes straight to storage without passing through your server. The moment it lands, Playbook tags it across 16+ dimensions (subject, color palette, style, usage rights, and more) and generates previews. No queue to run, no vision model to host, no thumbnail worker to keep alive. Your upload endpoint is finished the instant that POST returns.
Buyers search the way they describe things
Nobody browsing a stock library remembers a filename. GET /v1/{ORG}/ai_search runs a hybrid query: semantic embeddings on what an image or video actually looks like, plus full-text on titles, tags, and metadata, merged into one ranked list. Ask for "golden hour rooftop, muted tones," pass filters[media_type] or filters[boards][], and page with a cursor. It is the same GPT-powered search that runs inside Playbook, exposed as an endpoint. Talk to your creative files, then hand the results straight to your buyers.
Previews to everyone, originals to buyers
The paywall is the difference between a display_url and a raw_url. Every asset carries a CDN display_url for public browsing and a raw_url for the licensed original. Show previews freely. On purchase, call POST /v1/{ORG}/assets/{assetToken}/share to mint a link, then GET /v1/shared/{ORG}/assets/{assetToken}/download to resolve the buyer's download. Attach a license PDF that grants a named buyer usage for a set date range, and issue content credentials, invisible provenance via the Mentaport partner, so authorship travels with the file. That is your image licensing API, sitting in the same call graph as delivery.
What powers it
POST /v1/{ORG}/assets/upload_prepare: signed-URL ingest, files go straight to storage.- AI auto-tagging on upload: 16+ dimensions, no model for you to host.
GET /v1/{ORG}/ai_search: hybrid semantic plus full-text search, with filters and cursor pagination.- Custom fields: attach your own structured metadata (price, SKU, collection) and filter on it.
- CDN
display_urlandraw_url: resizable previews and licensed originals, permalinks included. POST .../shareandGET .../download: gated delivery resolved from a shared slug.- Licensing and rights: license PDFs, expiration, and content credentials via Mentaport.
- Bearer token auth on every call.
Two references worth bookmarking: the Playbook REST API docs and a worked marketplace example that builds exactly this flow, endpoint by endpoint.
The proof is live. The Vault Stock runs a creative stock platform on Playbook: 50,000+ assets, 100+ new uploads a week, and 3,000+ monthly members, with AI tagging, color and content search, and scheduled publishing. "Playbook really ticks all the boxes for us," says Anna Hamilton, Creative 2IC.
Developer FAQ
Which plans include the API? API access is a paid-plan feature, with monthly call quotas that scale by tier and SDK access on the higher tiers. Plans and limits change, so confirm the current numbers on the pricing page at the time of writing.
Do buyers need a Playbook account? No. Search, previews, and delivery run server to server with your bearer token. Buyers only ever touch your storefront and the download link you hand them.
Is search semantic or keyword? Both. One call merges embedding-based visual search with full-text matching, so "moody, backlit, vertical" and an exact tag both resolve in the same result set.
When are new uploads searchable? After background tagging and vectorization finish. A just-uploaded asset is stored instantly and becomes searchable once indexing completes, usually within seconds to minutes.
Point your storefront at a backend that already exists
You could spend two quarters building ingest, a vision tagger, a search index, a CDN, and a licensing system. Or you could point your asset marketplace platform at an API that already does all of it. Playbook brings your work to life. You just build the front door.
Ready to see the digital asset marketplace API behind your next build? Schedule a demo.