Kling AI — video

Text-to-video and image-to-video via Kling AI. Models: kling-v1-6 / kling-2.6 / kling-v3 / kling-v2-1-master / kling-v2-5-turbo. Async, typical time 2–5 minutes.

The integration gotcha

Fields mode, resolution and duration_seconds are REQUIRED. Skip any of them or pass a value off the list and you'll get 422 "Pricing is not configured for this bundle". The bundle key is hashed from (provider, model, method, mode, resolution, duration) — without an exact match against a seeded combo, the bundle is created with no price.

Parameters

Name	Type	Required	Description
prompt	string	✓	Scene description, non-empty.
mode	enum	✓	`standard` \| `pro` — standard = 720p (cheaper), pro = 1080p (pricier, better).
resolution	enum	✓	`720p` \| `1080p` — Locked to mode: standard ↔ 720p, pro ↔ 1080p.
duration_seconds	integer	✓	`5` \| `10` — Clip length in seconds. PER_SECOND pricing — cost scales linearly.
input_images	string[]	✓ for i2v	1–2 public URLs for image_to_video. https:// only.
aspect_ratio	enum	—	`16:9` \| `9:16` \| `1:1`

Valid mode × resolution × duration combinations

Anything off this matrix triggers price_not_configured. Prices are seeded ONLY for these combos.

mode	resolution	duration_seconds
standard	720p	5, 10
pro	1080p	5, 10

docs.klingMatrixNote

Text-to-video

docs.klingT2vBody

curl -X POST https://api.aigenway.com/v1/generations \
  -H "Authorization: Bearer sk_live_..." \
  -H "Idempotency-Key: $(uuidgen)" \
  -H "Content-Type: application/json" \
  -d '{
    "provider": "kling_ai",
    "model": "kling-v3",
    "method": "text_to_video",
    "params": {
      "prompt": "A red sports car drifting around a mountain curve at sunset, cinematic",
      "mode": "pro",
      "resolution": "1080p",
      "duration_seconds": 5,
      "aspect_ratio": "16:9"
    }
  }'

Image-to-video

Attach 1–2 public image URLs in input_images. The first-frame must be a publicly reachable https URL.

curl -X POST https://api.aigenway.com/v1/generations \
  -H "Authorization: Bearer sk_live_..." \
  -H "Idempotency-Key: $(uuidgen)" \
  -H "Content-Type: application/json" \
  -d '{
    "provider": "kling_ai",
    "model": "kling-v3",
    "method": "image_to_video",
    "params": {
      "prompt": "Subject turns their head slightly, light wind in their hair",
      "input_images": ["https://your-cdn.example.com/portrait.jpg"],
      "mode": "pro",
      "resolution": "1080p",
      "duration_seconds": 5
    }
  }'

Polling for status and result

docs.klingPollingBody

# After POST returns { "task_id": "task_..." }, poll the task endpoint:
curl https://api.aigenway.com/v1/tasks/<task_id> \
  -H "Authorization: Bearer sk_live_..."

# Or the alias (equivalent, returns the same shape):
curl https://api.aigenway.com/v1/generations/<task_id> \
  -H "Authorization: Bearer sk_live_..."

# When status == "SUCCEEDED", fetch the result files:
curl https://api.aigenway.com/v1/tasks/<task_id>/result \
  -H "Authorization: Bearer sk_live_..."

Available models

All models support text_to_video + image_to_video. See the catalog for exact prices and supported mode/resolution combos.

kling-v1-6 — Legacy, cheapest (~$0.14 / $0.49 std/pro for 5s)
kling-2.6 — Balanced quality and price (~$0.56 / $1.12)
kling-v3 — Latest generation, native audio in pro (~$0.42 / $0.56)
kling-v2-1-master — Cinematic, pro only (~$0.70)
kling-v2-5-turbo — Fast high quality, pro only (~$0.45)