Content and video

All four tools run as jobs: a 202 answer, the result by polling or webhook (Jobs and results). Voice, music, closing phrase and avatar are picked from the reference lists (Reference data).

Content pack

POST /content/generate - scope tools:contentgen. From property facts and photos GREM makes descriptions, social posts, a narrated video, a landing page and a PDF - in one job. A draft listing is created for the pack.

At least 4 photos are required - as public addresses in form.media.photos. Upload your own files with POST /uploads first and pass the returned urls (Files).

No user field is needed: the pack is created on behalf of the key's account.

curl -X POST https://developers.grem.capital/api/v1/content/generate \
  -H "Authorization: Bearer gsk_live_xxx" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: 0c0e3a5b-7d1e-4b8a-9b2f-1e6d4c3a2b10" \
  -d '{
    "form": {
      "language": "eng",
      "propertyCategory": "apartment", "dealType": "sale",
      "title": "Bright 2-room flat near the park",
      "totalArea": 64, "rooms": 2, "price": 320000, "priceCurrency": "EUR",
      "addressCountry": "Germany", "addressCity": "Berlin",
      "media": { "photos": ["https://example.com/1.jpg", "https://example.com/2.jpg",
                             "https://example.com/3.jpg", "https://example.com/4.jpg"] }
    }
  }'

202 response: jobId (job id) and propertyId (the draft listing). The finished job (GET /content/generate/jobs/{id}) holds:

File: GET /artifacts/v1/ai/assistant/announcement/jobs/{jobId}/artifacts/{artifactId}/download. Events - content.completed, content.partial (some files failed), content.failed. Fewer than 4 photos - 400.

Listing texts

POST /text/generate - scope tools:text. A description, a presentation and a social post. Needs either objectId or at least one of description, city, totalArea - otherwise 422 with code: not_enough_input.

curl -X POST https://developers.grem.capital/api/v1/text/generate \
  -H "Authorization: Bearer gsk_live_xxx" \
  -H "Content-Type: application/json" \
  -d '{"city":"Berlin","totalArea":64,"rooms":2,"propertyCategory":"apartment","language":"en","description":"Renovated, quiet courtyard, 5 min to U-Bahn."}'

202 with jobId. The texts are in the job files, as with the content pack; the job also carries progress, masterProfile and error. Events - text.completed / text.failed.

Video from photos

POST /video/create - scope tools:video. A short narrated video from your photos or a listing's photos.

Body - multipart/form-data, each option as its own form field (not JSON in a data field):

curl -X POST https://developers.grem.capital/api/v1/video/create \
  -H "Authorization: Bearer gsk_live_xxx" \
  -F "[email protected]" -F "[email protected]" -F "[email protected]" \
  -F "prompt=Sunny 2-room flat in Berlin Mitte, quiet courtyard" \
  -F "length=short" -F "aspect=9:16" -F "language=en"

202 with jobId; the video is in the job files. The job also carries progress, masterProfile and error, as with the content pack. The price depends on length. Events - video.completed / video.failed.

Avatar video

POST /avatar/video - scope tools:avatar. A presenter talks about your listing.

curl -X POST https://developers.grem.capital/api/v1/avatar/video \
  -H "Authorization: Bearer gsk_live_xxx" \
  -H "Content-Type: application/json" \
  -d '{"avatarId":"AVATAR_ID","objectId":"OBJECT_ID","voiceId":"VOICE_ID","scriptText":"..."}'

202 with jobId. Video: GET /artifacts/v1/ai/assistant/avatar/jobs/{jobId}/artifacts/{artifactId}/download, artifactId from the artifacts of the finished job. Events - avatar.completed / avatar.failed. Missing avatarId, objectId or a short text - 400 with the fields in errors.

Related

• Reference data - voices, avatars, music, closing phrases.
• Jobs and results - states and files.
• Files - uploading photos.