Each entry below documents a dated API version. Use the RawUGC-Version header or pin your API key to select a specific version. See Versioning for details.
2026-03-06
Semantic resource IDs, unified media model, and REST route cleanup.
What’s new
- Semantic resource IDs — all resources now use prefixed IDs:
vid_xxx for videos, img_xxx for images, mus_xxx for music. These replace the old taskId and id fields.
- Consistent response fields — generation endpoints return
videoId, imageId, or musicId. Status endpoints return url (was resultUrl).
versions array on GET /videos/{videoId} — edit history showing all versions for a video (original + captions/overlays). Each entry includes videoId, version, url, operation, and createdAt.POST /videos/generate is the canonical video generation endpoint./videos/{videoId}/captions and /videos/{videoId}/overlay — editing endpoints return an EditResponse with videoId, url, version, operation, creditsUsed.RawUGC-Version response header on all API responses.
Breaking changes from 2025-11-10
| Before | After |
|---|
taskId in responses | videoId / imageId / musicId |
id + taskId on videos | videoId only |
resultUrl | url |
imageUrl (music album art) | albumArtUrl |
GET /tasks/{id} | Removed — use typed endpoints |
GET /images/{taskId} | GET /images/{imageId} |
GET /music/{taskId} | GET /music/{musicId} |
How to migrate
Set RawUGC-Version: 2026-03-06 in your request headers to opt in. Legacy IDs (raw database IDs and provider task IDs) are still accepted as URL parameters for backward compatibility.
# Generate a video
curl -X POST https://rawugc.com/api/v1/videos/generate \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "RawUGC-Version: 2026-03-06" \
-H "Content-Type: application/json" \
-d '{"model": "sora-2-text-to-video", "prompt": "A serene beach"}'
# Response: { "videoId": "vid_xxx", ... }
# Poll status using the videoId
curl https://rawugc.com/api/v1/videos/vid_xxx \
-H "Authorization: Bearer YOUR_API_KEY"
2025-11-10
Initial public API version.
What’s included
- Video generation — Sora 2 (text-to-video, image-to-video), Kling 2.6 Motion Control, Veo 3 / Veo 3 Fast
- Image generation — Nano Banana models (text-to-image, image editing)
- Music generation — Suno models (V3.5 through V5)
- Social media scheduling — TikTok, Instagram, YouTube post scheduling
- Content management — Personas, products, styles, messaging templates
- Viral library — Search and analyze TikTok videos with AI-powered insights
- Research tools — TikTok scraping, video analysis, content plans
This is the baseline version — no migration needed. 
