Localize an email template over the API

POST /documents with content_type: "html" and your target languages. The HTML goes in verbatim; formatting, links, and template placeholders all survive translation.

Then estimate and run exactly as in Automate Transept with the API: the estimate is free, and the run fans across every language version in one call.

# BASE="https://app.transept.ai/api/public/v1"; KEY from Settings → Developer
curl -s -X POST $BASE/documents \
  -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"title": "Trial welcome", "content_type": "html", "source_language": "en",
       "target_languages": ["de", "fr"],
       "content": "<h1>Welcome, {{ first_name | default:\"friend\" }}!</h1><p>Your trial starts today.</p>"}'
# → { document_id, language_group: [ {id, target_language}, … ] }

{{ first_name | … }} is masked before the model sees the text and restored verbatim after; the reply is counted against the source, so a lost or garbled token is repaired rather than shipped. One deliberate exception: fallback text inside a tag (the friend in default:"friend") is what recipients without a name on file actually read, so it is translated while the tag around it is not. The full token-protection rules are in strings files.

GET /documents/{id}/content?format=html on each language version returns the same HTML shape with translated copy, ready to paste into the ESP's language variant.

curl -s "$BASE/documents/<de-version-id>/content?format=html" \
  -H "Authorization: Bearer $KEY"
# → the same HTML with German copy, Liquid intact

Yes. A webhook trigger set to “the request body is the content” takes the template straight from your ESP or a Zapier/n8n step, runs the workflow on arrival, and an outgoing webhook pushes a signed event when the translations are ready, so your automation can pull the HTML back without polling. Nobody uploads anything by hand.