Documentation Index
Fetch the complete documentation index at: https://runcrate.ai/docs/llms.txt
Use this file to discover all available pages before exploring further.
Edit existing images using natural language. Change colors, swap backgrounds, remove objects, apply style transfer — one endpoint, no masking.
Available models
| Model | Strengths |
|---|
| FLUX.1 Kontext Pro | Fast edits, good instruction following |
| FLUX.1 Kontext Max | Highest quality, production assets |
| Wan2.6 Image Edit | Object removal, background replacement |
| Qwen Image Edit | Color/material changes, CJK text rendering |
Basic image edit
from runcrate import Runcrate
client = Runcrate(api_key="rc_live_YOUR_API_KEY")
edited = client.models.generate_image(
model="black-forest-labs/FLUX.1-kontext-pro",
prompt="Change the car color to matte black",
image="./red-car.png",
)
edited.data[0].save("black-car.png")
image accepts a file path (auto base64-encoded), a URL, or raw base64 data.
Background removal
from runcrate import Runcrate
client = Runcrate(api_key="rc_live_YOUR_API_KEY")
result = client.models.generate_image(
model="Wan-AI/Wan2.6-Image-Edit",
prompt="Remove the background, replace with clean white studio backdrop",
image="./product-on-table.png",
)
result.data[0].save("product-isolated.png")
Style transfer
result = client.models.generate_image(
model="black-forest-labs/FLUX.1-kontext-max",
prompt="Convert to a watercolor painting style with soft pastel colors",
image="./city-photo.jpg",
)
result.data[0].save("city-watercolor.png")
Batch editing
from runcrate import Runcrate
from pathlib import Path
from concurrent.futures import ThreadPoolExecutor
client = Runcrate(api_key="rc_live_YOUR_API_KEY")
output = Path("./clean"); output.mkdir(exist_ok=True)
def clean(p: Path):
r = client.models.generate_image(model="Qwen/Qwen-Image-Edit", prompt="Remove background, white backdrop", image=str(p))
r.data[0].save(str(output / p.name))
with ThreadPoolExecutor(max_workers=4) as pool:
pool.map(clean, list(Path("./raw").glob("*.png")))
Tips
- Kontext Pro vs Max. Pro for fast iteration, Max for final production assets.
- Be specific. “Change the sky to vivid cerulean blue with wispy cirrus clouds” beats “make the sky bluer.”
- Only describe the change. The models preserve the subject — describe only what should change.