ironfang-renderwolf

The official Python client for the Renderwolf API: screenshots, PDFs, QR codes, template images, clips and site previews from a URL or from HTML you send. No dependencies; Python 3.10 and later.

pip install ironfang-renderwolf

import os
from ironfang_renderwolf import Renderwolf

with Renderwolf(api_key=os.environ["IRONFANG_API_KEY"]) as rw:
    shot = rw.screenshot(url="https://example.com", full_page=True)
    shot.save("page.png")
    print(shot.credits, shot.render_ms, shot.request_id)

Every render returns the bytes plus what the API reports about them: content type, credits charged (zero on a cache hit), render time and the request id to quote to support. Long outputs stream: save() and iter_bytes() never hold a whole clip in memory.

Durable jobs

job = rw.submit_job("pdf", {"html": invoice_html, "paper_format": "a4"}, idempotency_key=f"invoice-{invoice.id}")
done = rw.wait_for_job(job["id"])
if done["status"] == "succeeded":
    rw.job_result(job["id"]).save(f"invoice-{invoice.id}.pdf")

Verifying a webhook

Run the check over the raw request body, before parsing it.

from ironfang_renderwolf import verify_webhook

ok = verify_webhook(
    secret=os.environ["RENDERWOLF_WEBHOOK_SECRET"],
    timestamp=request.headers["Renderwolf-Timestamp"],
    signature=request.headers["Renderwolf-Signature"],
    payload=request.get_data(),
)

Errors

Failures raise RenderwolfError with status, the API's code (quota_exhausted, render_failed, ...) and request_id. Synchronous renders are never retried by the client: a render that timed out may already have been charged, and retrying it blind doubles the bill.