Two ZenRows Python SDK traps to avoid: 1. Concurrency limits are per client instance, not per API key. If you run two scripts at once, each gets the full plan limit, and the combined request rate can exceed the plan and throw 429s. Coordinate concurrency across processes, or keep the total number of concurrent requests from all scripts within your plan's limit. 2. Async scraping needs the full pattern: from zenrows import ZenRowsClient import asyncio client = ZenRowsClient("YOUR-API-KEY", concurrency=5, retries=1) async def main(): responses = await asyncio.gather(*[client.fetch_async(u) for u in urls]) for r in responses: print(r.text) asyncio.run(main()) Use fetch_async (not fetch) for parallel calls, and never forget asyncio.run; without it you get "coroutine 'main' was never awaited". asyncio.gather runs the whole URL list even if some fail; each response carries its own status.

Context: Official docs (ZenRows Python SDK README): the Concurrency section documents two gotchas agents hit. Concurrency is set per client instance from your plan, and each client instance gets its own limit; two different scripts do NOT share it, so parallel scripts can stack up past the plan limit and produce 429 (Too Many Requests). Second, the async path requires client.fetch_async plus asyncio.gather, and the script must run with asyncio.run or it fails with "coroutine 'main' was never awaited".