snapedit-api 1.0.0

Official Python SDK for the SnapEdit API — 40+ AI image & video editing models through a single API key.

SnapEdit Python SDK

Official Python SDK for the SnapEdit API — 40+ AI image & video editing models through a single API key. Sync and async, fully typed.

Install

pip install snapedit-api

Requires Python 3.8+.

Quick start

from snapedit import SnapEdit

client = SnapEdit(api_key="sk-snap-...")  # or set SNAPEDIT_API_KEY

out = client.remove.background("https://example.com/photo.jpg")
print(out.data[0].url)

Inputs accept a URL (fetched server-side), a local file path, bytes, a pathlib.Path, or any binary file-like object:

client.enhance.image("./portrait.jpg", zoom_factor=4, enhance_faces=True)

with open("photo.png", "rb") as f:
    client.remove.background(f)

Async

import asyncio
from snapedit import AsyncSnapEdit

async def main():
    async with AsyncSnapEdit() as client:  # reads SNAPEDIT_API_KEY
        out = await client.generate.zimage("a red panda astronaut", aspect_ratio="16:9")
        print(out.data[0].url)

asyncio.run(main())

What's available

Group	Methods (examples)
client.detect	objects, text, wires
client.remove	background, objects, text, logo, wires, reflection, mirror, background_graphic
client.enhance	image, image_pro, art, restore, colorize, light_restore, backlit_fix, night_flash
client.generate	zimage, qwen, art, background, headshot, sticker
client.edit	edit, multi
client.beauty	retouch_skin, makeup, hairstyle
client.utility	outpaint, pose_suggest, ai_detect
client.try_on	create, get, cancel, wait, run (async task)
client.video	enhance, remove_logo (upload-based async task)
client.files	remove_logo_pdf (upload-based async task)

Async task endpoints

Virtual try-on, video, and PDF endpoints are long-running. Use the one-shot run / enhance / remove_logo helpers to submit and poll to completion:

task = client.try_on.run(
    model_image="https://example.com/model.jpg",
    cloth_image="https://example.com/shirt.jpg",
    cloth_type="upper",
)
print(task.download_url)

# Videos upload the file for you, then poll:
result = client.video.enhance("./clip.mp4", zoom_factor="2K")
print(result.download_url)

Control polling with WaitOptions:

from snapedit import WaitOptions

task = client.try_on.run(
    model_image="...", cloth_image="...", cloth_type="upper",
    opts=WaitOptions(poll_interval=2, max_wait=600,
                     on_progress=lambda t: print(t.status, t.progress)),
)

Configuration

client = SnapEdit(
    api_key="sk-snap-...",
    auth_scheme="bearer",   # or "api-key" header
    timeout=60.0,           # per-request seconds
    max_retries=2,          # auto-retry 429 / 5xx / network errors
    base_url="https://api.snapedit.app",
)

Errors

from snapedit import SnapEditError, SnapEditTaskError

try:
    client.remove.background("bad-input")
except SnapEditError as e:
    print(e.status, e.type, e.code)
    if e.is_rate_limit:
        print("retry after", e.retry_after, "seconds")

SnapEditTaskError is raised when an async task fails or a wait() times out.

Rate limits

After any request, inspect client.rate_limit (limit, remaining, reset).

License

MIT