mediamart for humans and agents

One flat media library. Drop a file, it is auto-tagged and re-rendered for every platform surface. Originals are never modified.

Start in three calls

Mint a key in your account.
POST /v1/assets with the name, content type and byte count.
PUT the bytes to the returned uploadUrl, then POST /v1/assets/{id}/finalize.

curl -X POST https://api.mediamart.555.team/v1/assets \
  -H "Authorization: Bearer mm_<your key>" \
  -H "content-type: application/json" \
  -d '{"name":"poster.jpg","contentType":"image/jpeg","bytes":482113}'

Analysis and rendering run in the background; poll GET /v1/assets/{id} until status is ready.

Agents

The same key opens an MCP server at POST /v1/mcp — stateless JSON-RPC 2.0, four tools: mediamart_search, mediamart_get, mediamart_plan, mediamart_upload.

curl -X POST https://api.mediamart.555.team/v1/mcp \
  -H "Authorization: Bearer mm_<your key>" \
  -H "content-type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

llms.txt is the one-page brief for a model; the reference is the full surface.

What each call answers

Endpoint	Purpose
`POST /v1/assets`	Reserve an upload, get a presigned PUT
`POST /v1/assets/{id}/finalize`	Record the object, enqueue the pipeline
`GET /v1/assets`	Search by word, tag or platform
`GET /v1/assets/suggest`	Tag and title type-ahead
`GET /v1/assets/{id}`	Caption, tags, every derivative
`POST /v1/plan`	Crop plan for a source size, no upload
`GET /v1/usage`	Analyses, exports and storage against the plan