Card Trick Car Reveal

Flow·3 steps

Upload your photo and any car. One card trick, and the car of your dreams is parked right behind you.

Steps
3
Inputs
2
Runs
30
Template · pipeline3 steps
  1. 1
    gpt-image-v2-5-sunburst-editid · step1 · model
    background
    "auto"
    image_size
    "1536x1024"
    image_urls
    [1 item]
    moderation
    "low"
    num_images
    1
    …
    +3 more params
  2. 2
    bytedance-seedance-2-5-reference-to-videoid · step2 · model
    aspect_ratio
    "auto"
    duration
    "auto"
    generate_audio
    false
    image_urls
    [2 items]
    prompt
    " @[Video 1](video_1) = ORIGINAL VIDEO: source of motion, tim…"
    …
    +2 more params
  3. 3
    audio-replacerid · step3 · model
    mode
    "replace"
    new_audio
    "https://cdn-us.eachlabs.ai/uploads/31807d1f-84c7-4284-9995-c…"
    video
    "$.step2.primary"
How to use this template
trigger.sh
sh
# 1. Trigger the workflow
EXEC=$(curl -s -X POST https://api.eachlabs.ai/v1/workflows/trigger/11eaeb85-4272-41fd-aa55-233cee2ac353/v1 \
  -H "Authorization: Bearer $EACHLABS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
      "inputs": {
        "Car": "https://cdn-us.eachlabs.ai/defaults/16df7c7e18164d398065a3275ff98a9c.png",
        "Person": "https://cdn-us.eachlabs.ai/defaults/efeed4ec80da4b67a8f8461b38472213.png"
      }
    }' | jq -r .execution_id)

# 2. Poll the execution until it's done
curl -s https://api.eachlabs.ai/v1/workflows/executions/$EXEC \
  -H "Authorization: Bearer $EACHLABS_API_KEY"

Trigger returns an execution_id · poll the executions endpoint for the final output

Playground · what you can change2 inputs

Clone this template into your workspace, then tweak any of the inputs below, or rewrite the step prompts, to spin up your own variation. The pipeline stays the same; the creative output is yours.

  • Car
    default · 16df7c7e18164d398065a3275ff98a9c.png
    image URL
  • Person
    default · efeed4ec80da4b67a8f8461b38472213.png
    image URL
Ideas for your remix
  • · Swap the input media to retell the story with different subjects.
  • · Edit the prompt inside gpt-image-v2-5-sunburst-edit to change tone, scene, or pacing.
  • · Replace any model step with a faster or pricier alternative for a different vibe.
  • · Chain extra steps (upscale, audio, captions) to extend the pipeline.

More flows like this