How do I get past WeChat's 200013 freq control error when fetching Official Account articles?

You start a backfill of 40 Official Accounts, it runs fine for twenty minutes, and then every request answers ret=200013, freq control. From then on nothing works until the next day. The fix that lasts is to stop sending those requests from your own WeChat account. Through Official Account article history you page through each account's articles with an API key; there is no backend login on your side for WeChat to throttle.

What 200013 actually means

Scrapers that read article lists from mp.weixin.qq.com sign in with an Official Account you own and call the backend's article search as that account. WeChat counts those calls per account. Past a threshold it returns 200013 and blocks that function for hours. Changing your IP doesn't help because the limit sits on the account. The usual workarounds, longer sleeps between requests and several backend accounts in rotation, make a backfill take days and add more logins that expire.

The backfill, rewritten

Each history call returns 10 articles and a cursor for the next page. Pace calls to the rate limit of your key and retry the errors that say they can be retried:

import os, time, uuid, requests

API = "https://api.okdata.online"
H = {"Authorization": f"Bearer {os.environ['OKDATA_API_KEY']}"}

def call(path, **params):
    key = str(uuid.uuid4())  # same key on every retry: never pay twice
    for _ in range(5):
        r = requests.get(API + path, params=params, timeout=90,
                         headers={**H, "Idempotency-Key": key})
        body = r.json()
        if body["code"] == "ok":
            return body
        if body.get("retry") in ("yes", "same_key"):
            time.sleep(body.get("retry_after_seconds") or 5)
            continue
        raise RuntimeError(f"{body['code']}: {body['message']}")
    raise RuntimeError("gave up after 5 tries")

def history(gh):
    cursor = None
    while True:
        body = call("/v1/wechat_mp/accounts/articles/list", id=gh, **({"cursor": cursor} if cursor else {}))
        yield from body["data"]["items"]
        if not body["page"]["has_more"]:
            return
        cursor = body["page"]["next_cursor"]

What a backfill costs

Job	Calls	Cost
1 account, 300 articles, list only	30	$1.20
40 accounts, 300 articles each, list only	1,200	$48.00
Same, plus full text of 2,000 of them	1,200 + 2,000	{{cost:wechat_mp.accounts.articles.list1200+wechat_mp.articles.get2000}}

Errors are free, including the busy and timeout cases the loop above retries. Full text comes from article detail.

Limits that still apply

• Your key has its own rate limit. Going over it returns rate_limited with a wait time; the X-RateLimit-Remaining header tells you how close you are. The error reference lists every code.
• If the data source is busy you get upstream_busy, not charged, retry shortly.
• Only public articles. Reads and likes are filled when available and null otherwise.

If you'd rather keep scraping yourself

Space requests several seconds apart, keep each backend account's daily volume modest, and stop for the day on the first 200013 instead of pushing on, which tends to lengthen the block. That works for a single account you archive slowly. For your own account, WeChat's official API returns your published articles without going through the backend search at all. For dozens of accounts on a deadline, an API is the quicker road; see competitor monitoring.