Quickstart

From zero to a generated, ESPR-ready Digital Product Passport in a handful of cURL commands - then hand the same key to an AI agent so it keeps producing them for you. No package install, no language client; every example is plain HTTP against the REST API.

1. Create a workspace

Sign up and a personal workspace is provisioned for you automatically. A workspace owns your passports, learnings, settings and API keys, and carries the AI mode (Review or Auto). You can rename it and invite teammates later from Dashboard > Workspace.

2. Create a master API key

Open Settings > API Keys in the dashboard and create a key. The secret (it starts with ep_live_sk_) is shown once - copy it now and export it:

# Settings > API Keys in the dashboard creates the key.
# Copy the secret (shown once) and export it for this shell.
export DPP_API_KEY="ep_live_sk_..."

• Pick the permissions the key needs: read, write, delete (an admin key grants all).
• Store keys in environment variables or a secrets manager - never in your source tree.
• If a key leaks, rotate or revoke it from the dashboard; the old key stops working right away.

Confirm the key works by reading your workspace:

curl https://dppautomate.com/api/v1/workspace \
  -H "Authorization: Bearer $DPP_API_KEY"

{
  "data": {
    "id": "...",
    "name": "Acme GmbH",
    "mode": "review",
    "role": "owner",
    "entitlement": {
      "tierId": "starter",
      "active": true,
      "capacity": 100,
      "activePassportCount": 12,
      "needsPlan": false
    }
  }
}

3. Generate a passport from an image

The fastest path to a passport: send a product photo to POST /api/v1/passports/generate. The image is a data URL or raw base64. The workspace AI mode decides what happens to the result - in Review mode it lands as a draft; in Auto mode it goes live. Pass create: falsefor a preview that isn't saved.

curl https://dppautomate.com/api/v1/passports/generate \
  -X POST \
  -H "Authorization: Bearer $DPP_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "image": "data:image/png;base64,<BASE64_OF_PRODUCT_PHOTO>"
  }'

The response wraps the passport with the mode and resulting status:

{
  "data": {
    "passport": {
      "id": "...",
      "object": "digital_product_passport",
      "identification": {
        "productName": "Merino Wool Beanie",
        "productCategory": "Textiles & Apparel",
        "dppUid": "DPP-1717-AB12CD"
      },
      "status": "draft"
    },
    "mode": "review",
    "status": "draft",
    "created": true
  }
}

4. List passports

Read your passports back with pagination and filters (status, category, search, favorite, per_page):

curl "https://dppautomate.com/api/v1/passports?status=active&per_page=25" \
  -H "Authorization: Bearer $DPP_API_KEY"

5. Approve a draft

In Review mode, generations queue up as drafts. List them, then approve the ones you trust - approval promotes the draft to an active passport and captures any human edits as learnings.

# 1. List drafts awaiting review.
curl https://dppautomate.com/api/v1/drafts \
  -H "Authorization: Bearer $DPP_API_KEY"

# 2. Approve one -> it becomes an active passport.
curl -X POST https://dppautomate.com/api/v1/drafts/PASSPORT_ID/approve \
  -H "Authorization: Bearer $DPP_API_KEY"

6. Add a learning

Learnings steer future generations. Add one whenever you notice the AI getting a field wrong; active learnings are injected into the prompt so the next passport is more accurate.

curl https://dppautomate.com/api/v1/learnings \
  -X POST \
  -H "Authorization: Bearer $DPP_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "instruction": "Always read the country of origin from the sewn-in label.",
    "section": "production"
  }'

7. Set the mode

Once the model is dialled in for your catalogue, switch the workspace to Auto so new passports go live without a manual approval step:

curl https://dppautomate.com/api/v1/workspace/mode \
  -X PUT \
  -H "Authorization: Bearer $DPP_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "mode": "auto" }'

8. Hand it to an AI agent

DPPAutomate is built to be operated by an LLM. Point your agent at the machine-readable spec and the operator guide, give it the master key, and it can run the whole loop - generate, review, learn, publish. A prompt like this is enough to get started:

You manage my DPPAutomate workspace via its REST API.

- Spec:        https://dppautomate.com/api/openapi.json
- Operator guide: https://dppautomate.com/llms.txt
- Auth: send "Authorization: Bearer $DPP_API_KEY" on every /api/v1 request.

Task: for each product photo I give you, call
POST /api/v1/passports/generate. We are in Review mode, so each one
becomes a draft - list them with GET /api/v1/drafts and summarise what
needs a human decision. When I correct a field, add a learning via
POST /api/v1/learnings so future passports improve.

The two URLs the agent needs: /api/openapi.json (the full OpenAPI 3.1 contract) and /llms.txt (a concise operator guide). For the full endpoint list, permissions and error model, see the API reference.