Autom

Python

Install the Autom Python SDK and make your first API call.

The official Python SDK for Autom, with one typed method per endpoint and a generic request() for anything else.

Requirements

  • Python 3.11+

Install

pip install automdev

Get your API key

Copy your API key from the Autom dashboard and set it in the AUTOM_API_KEY environment variable, which the SDK reads by default. Keep it out of source control.

Quick start

import os

from autom import AutomClient

with AutomClient(os.environ["AUTOM_API_KEY"]) as client:
    response = client.google_search(query="web scraping api")
    print(response.data)

Async jobs

Any call accepts is_async, which is useful past the 60-second server limit. The response is then a job_id, and the SDK can wait for the job to finish.

submitted = client.google_search(query="web scraping api", is_async=True)
job = client.wait_for_job(
    submitted.data["job_id"],
    on_progress=lambda j: print(j["status"], j.get("progress")),
)
print(job.data)

Errors

A non-2xx response raises an API error with its status and message. See Status codes for the full list.

from autom import AutomAPIError

try:
    client.google_search(query="web scraping api")
except AutomAPIError as error:
    print(error.status, error, error.response_body)

Retries

Requests are retried up to 3 times after a network error, a 408, 500, 502, 503 or 504, or a plan rate limit. Per-key quota errors and client-side timeouts are not retried. The delay starts at 1 second and grows 1.5× per retry, and a Retry-After header wins.

client = AutomClient(max_retries=5, retry_delay=0.5)

# or for one request
client.google_search(query="web scraping api", request_options=RequestOptions(max_retries=0))

Batch requests

Run many calls at once, 10 at a time by default. One failure does not stop the others.

inputs = ["web scraping api", "serp api", "rank tracker"]
results = client.batch(inputs, lambda query: client.google_search(query=query), concurrency=5)

for result in results:
    if result.ok:
        print(result.value.data)
    else:
        print(inputs[result.index], result.error)

Any endpoint

request() reaches any path, including endpoints released after your SDK version.

response = client.request("GET", "/v1/usage")

Common mistakes

  • Putting the API key in front-end code. Anyone could read it. Call the API from your server.
  • Treating every 401 as a bad key. A 401 can also be a rate limit: read the message. The SDK retries plan rate limits for you.
  • Retrying a 402. It means there are no credits or quota left, so a retry cannot succeed. See Credits.
  • Waiting on long requests. Past 60 seconds, use async mode as shown above rather than raising your timeout. See Timeout.

On this page

RequirementsInstallGet your API keyQuick startAsync jobsErrorsRetriesBatch requestsAny endpointCommon mistakesLinks