Kwaivgi Kling V2.0 I2V Master API Documentation

Kwaivgi Kling V2.0 I2V Master API Documentation

Playground

Kling 2.0 master elevates image-to-video with improved prompts, richer character motion, better visuals and a Multi-Elements Editor. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

Kling v2.0 I2V Master generates short, cinematic videos from a reference image and a motion-focused prompt. Upload a starting image, optionally provide an last_image for better continuity, and describe the action and camera movement to animate the scene with smooth motion and strong visual coherence. Built for stable production use with a ready-to-use REST API, no cold starts, and predictable pricing.

Key capabilities

• Image-to-video generation anchored to a reference image
• Optional last_image support for more controlled transitions and ending consistency
• Prompt-controlled motion: subtle gestures, facial micro-expressions, hair/clothing movement, environment effects
• Cinematic camera movement: push-in, pan, orbit, handheld micro-shake
• Optional negative_prompt to reduce artifacts and unwanted styles

Pricing

Inputs

• image (required): the starting reference image (visual anchor)
• last_image (optional): target end frame reference for smoother continuity
• prompt (required): describe what happens and how the camera moves
• negative_prompt (optional): describe what to avoid (blur, distortions, artifacts)

Parameters

• prompt: motion + scene direction for the clip
• negative_prompt: optional “avoid list”
• image: input image (upload or URL)
• last_image: optional end-frame image (upload or URL)
• guidance_scale: how strongly the result follows your prompt (lower = more natural, higher = more literal)
• duration: video length in seconds

Prompting guide (I2V with last_image)

For best control, write prompts like a director’s brief:

• Subject action: what the main subject does over time (smile, turn, wave, stand up, walk)
• Environment motion: wind in trees, dust, rain, fog, drifting particles
• Camera motion: slow push-in, orbit, gentle pan, steady handheld feel
• Continuity rules: keep identity, outfit, and scene layout consistent; if using last_image, ensure the motion “lands” on that end frame naturally

Example prompts

• A warm, serene park scene. An elderly couple sits on a bench; the wife gently pats her husband’s back, both smiling. Sunlight filters through leaves with soft dappled shadows. Camera slowly pushes in, subtle breeze moves their hair and clothing, natural motion, 5 seconds.
• Portrait outdoors at golden hour. The subject smiles and looks toward the camera, hair sways slightly in the wind. Gentle handheld micro-sway, shallow depth of field, 5 seconds.
• Cozy indoor scene. The subject turns a page in a book, warm lamp lighting, floating dust motes, camera slow orbit, 5 seconds.

Negative prompt examples

• blur, distort, low quality
• jitter, warping, flicker, melted details
• watermark, logo, subtitles, text artifacts

Authentication

For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result

set -euo pipefail

export WAVESPEED_API_KEY="your-api-key"

REQUEST_BODY=$(cat <<'JSON'
{
  "prompt": "A cinematic ocean wave at sunrise, highly detailed",
  "image": "https://interactive-examples.mdn.mozilla.net/media/cc0-images/painted-hand-298-332.jpg",
  "guidance_scale": 0.5,
  "duration": 5
}
JSON
)

# 1. Submit the prediction once. Never retry this request automatically:
# a retry after an unclear failure can create a second, billed task.
CHECK='The task may already exist. Check History or the predictions list before submitting again.'
if ! SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  --connect-timeout 10 --max-time 60 \
  -X POST "https://api.wavespeed.ai/api/v3/kwaivgi/kling-v2.0-i2v-master" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}"); then
  printf 'Submit failed. %s\n' "${CHECK}" >&2
  exit 1
fi

TASK=$(printf '%s' "${SUBMIT_RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
PREDICTION_ID=$(printf '%s' "${TASK}" | jq -r '.id // empty')
if [ -z "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id. %s\n' "${CHECK}" >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 2. Poll until the prediction reaches a final status.
BODY_FILE=$(mktemp)
trap 'rm -f "${BODY_FILE}"' EXIT
while true; do
  # Transient errors (429, 5xx) are retried, honoring Retry-After.
  HTTP_CODE=$(curl --silent --show-error --fail-with-body \
    --connect-timeout 10 --max-time 60 --retry 5 --retry-delay 2 \
    --output "${BODY_FILE}" --write-out '%{http_code}' \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}" || true)
  case "${HTTP_CODE}" in
    2??) ;;
    000|429|5??)  # dropped connection or temporary error: keep polling
      printf 'Task %s: HTTP %s, polling again in 10s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      sleep 10; continue ;;
    *) printf 'Task %s: HTTP %s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      cat "${BODY_FILE}" >&2; exit 1 ;;
  esac
  RESULT=$(jq 'if type == "object" and has("data") then .data else . end' "${BODY_FILE}")
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout|deleted)
      printf 'Task %s ended with status %s\n' "${PREDICTION_ID}" "${STATUS}" >&2
      printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    *) sleep 2 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Every model also accepts the global request options enable_sync_mode and enable_base64_output; see their pages for details.

Response Parameters

Result Request Parameters

Result Response Parameters