Documentation
Back
  • Getting Started
    • Create Your First Ad
    • Teams, Brands, and Navigation
    • Generation Basics
  • Brands
    • Create a Brand
    • Work in a Brand
    • Manage Brand Settings
    • Brand Guidelines and Website Data
    • Connect Publishing Accounts
    • Manage Brand Members
  • Create Ads
    • Playground vs Launchpad
    • Create Video Ads
    • Create Static Ads
    • Create UGC Ads and Personas
    • Create Rich Media Ads
    • Create Carousel Ads
    • Create Product Ads
  • Tools
    • Image Studio
    • Remove Background
    • Generate Vector
    • Adapt Ad
    • Video Generator
    • Edit Video
  • Campaigns & Personalization
    • Data Sheets
    • Personalize a Static Ad
    • Review and Approve a Set
    • Work with a Creative Set
    • Export a Set
    • AI Values per Row
    • Generate a Template from a Data Sheet
    • Personalized Rich Media Sets
    • Fit Long Values in Their Box
    • Sets in Campaigns and Assistants
  • Media & Delivery
  • Team & Account
    • Compliance Settings
  • Help & Troubleshooting
  • Developers & Automation
    • API
    • MCP
      • MCP setup

API

Use the AdGreetz Creative REST API to create video, static, carousel, and rich media ads from your own apps and backends.

AdGreetz exposes a versioned Creative REST API for server-to-server integrations. It uses the same generation pipelines as Creative Studio and returns JSON for every request.

https://app.adgreetz.com/api/v1

Authenticate with a bearer API key from Settings → API Keys, or a Supabase OAuth access token. Create keys and manage connected apps on that page.

Interactive reference

The full endpoint reference is published as OpenAPI 3.1 and rendered with Scalar:

  • Creative REST API reference — browse endpoints, schemas, and try requests in the browser
  • OpenAPI document — machine-readable spec for codegen and API tools

UGC ads

UGC uses the same video generation resource and editor as the UGC playground:

  1. Read GET /api/v1/ugc/options for models, supported durations and creator restrictions.
  2. Read GET /api/v1/ugc/creators?brand_id=BRAND_ID&source=approved_character. Show the returned names and preview_url images so the user can select a character. Use source=persona_image for that brand's persona images.
  3. Upload local product images with POST /api/v1/uploads and PUT the bytes before generating.
  4. Call POST /api/v1/ugc with an existing brand, the selected creator and product media:
{
  "brand_id": "BRAND_ID",
  "creator": { "type": "approved_character", "id": "asset-SELECTED" },
  "product_images": [{ "asset_id": "UPLOADED_PRODUCT_ASSET_ID" }],
  "duration": 15,
  "aspect_ratio": "9:16"
}

Creator alternatives are { "type": "persona_image", "asset_id": "IMAGE_ASSET_ID" } and { "type": "description", "text": "A detailed creator description" }. Persona photos cannot be used with Seedance; select a compatible model from the options response. Custom creators follow the workspace's persona entitlement.

The response contains generation: { type: "video", id }. Auto-render is enabled by default. Poll /api/v1/generations/video/ID using poll_after_seconds, then fetch /api/v1/generations/video/ID/result when rendered. Stop on failed, or on completed with an error: the composition is available but rendering needs recovery. For an edit-first workflow, pass auto_render: false and call the existing render endpoint when ready. Existing cancel and webhook conventions apply.

Submit generation directly. The workflow calculates costs and returns INSUFFICIENT_CREDITS if funds are insufficient. Tell the user and wait until they add funds and ask to continue. A queued job that later runs out of funds reports error_code: "INSUFFICIENT_CREDITS" on its status/webhook. Any returned credit estimate is informational; the agent does not need a balance precheck or cost calculation.

Use Idempotency-Key to retry a creation request safely within the existing API idempotency window. Product, scene, video and audio references accept an asset_id or a public HTTPS url; private library media must use asset IDs. Model-specific reference limits are enforced. API UGC currently supports team workspaces and existing brands; drafts/finalize and brand creation are excluded.

Related

For agent-first integrations over the Model Context Protocol, see the MCP docs and MCP setup guide.