FlashVSR Video Upscale API

Authentication

Authorization: Bearer YOUR_API_KEY

Model & pricing

⚠️ Every tier is priced per second of the INPUT video, not the output. For one 30-second source, the only thing that changes between 4K and 720p is the rate.

model = flashvsr

Code examples

# ─── Upscale a video (FlashVSR) ───────────────────────────
curl -X POST https://api.apimodels.app/v1/video/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "flashvsr",
    "video_url": "https://example.com/clip.mp4",
    "resolution": "1080p"
  }'

# Response (async — a task is created):
# { "code": 200, "msg": "success", "data": { "taskId": "clxxx", "state": "pending" } }


# ─── Poll for the upscaled clip ───────────────────────────
curl "https://api.apimodels.app/v1/video/generations?task_id=TASK_ID" \
  -H "Authorization: Bearer YOUR_API_KEY"
# { "code": 200, "data": { "state": "completed",
#     "resultUrls": ["https://r2.apimodels.app/videos/clxxx.mp4"] } }

Request parameters

Notes and limits

• Input must be an MP4 at a public URL, at most 120 seconds and 500 MB per clip. Requests beyond that are rejected before we submit them upstream, rather than failing minutes later.
• A URL whose duration we cannot read is rejected with that reason: billing is per source second, so an unmeasurable clip cannot be priced — and a URL we cannot fetch is one the upstream probably cannot fetch either.
• The output keeps the source duration and frame rate and only raises the resolution. Measured: a 320x180, 4-second, 24fps source came back as 1280x720, 4 seconds, 24fps in 25.5 seconds.
• Async: create a task to get a taskId, then poll GET ?task_id=. Terminal states are completed or failed; the finished clip is in data.resultUrls.
• Failed tasks are never billed.

Try it in the Playground