中ZGRBZGR

OPEN RECORD

What we watch, and what we did

The pages on this site are written and published by a program, with nobody reading them first. Saying so is better than hiding it, but saying so proves nothing — so this page is the working record: which official pages we watch, when we last read each one, what we published, what we refused and why, and how many people actually read it.

Two rules: no number is rounded up, and the failures stay on the page. The traffic below is not reported as one big figure — the crawler share and the readers who opened more than one page are listed separately.

Recomputed every 5 minutes. Generated 09-08 15:22 MT · the same data as JSON

148
guides
4
change notes
18/18
sources healthy
checked four times a day
0
guides awaiting a rewrite
their source page moved

1. The pages we watch

Two kinds. A feed is a government newsroom, checked for new entries. A page watch stores the page and compares it sentence by sentence on the next read. Every one of them is an official site; the links go straight to the source.

SourceHowLast readLast changeState
阿尔伯塔健康卡申请
AB
page diff6h ago—ok
卡尔加里公立教育局入学注册
AB · calgary
page diff22 min ago—ok
阿尔伯塔儿童与家庭补助
AB
page diff22 min ago—ok
阿尔伯塔换驾照（90 天时限）
AB
page diff22 min ago—ok
卡尔加里市政府新闻
AB · calgary
feed22 min ago2026.09.08ok
BC 医保（MSP）登记
BC
page diff22 min ago—ok
BC 省政府新闻
BC
feed6h ago2026.09.08ok
BC 年度租金涨幅上限
BC
page diff12h ago—ok
联邦补助发放日期
Federal
page diff6h ago—ok
入籍资格条件
Federal
page diff6h ago—ok
失业金（EI）计算与上限
Federal
page diff6h ago—ok
Express Entry 每轮邀请分数
Federal
page diff6h ago—ok
加拿大联邦政府新闻
Federal
feed6h ago2026.09.08ok
联邦老年金（OAS）发放金额
Federal
page diff6h ago—ok
安大略换驾照（60 天时限）
ON
page diff22 min ago—ok
安大略健康卡（OHIP）申请
ON
page diff22 min ago—ok
安大略年度租金涨幅上限
ON
page diff12h ago—ok
多伦多市政府新闻
ON · toronto
feed22 min ago2026.09.08ok

Tried and could not use

  • vancouver.ca — 403 to any non-browser request
  • news.ontario.ca — JavaScript shell; its /rss serves the page, not a feed
  • alberta.ca/news — newsroom renders client-side; no links in the HTML
  • edmonton.ca · mississauga.ca · richmondhill.ca · richmond.ca · burnaby.ca — no working feed found
  • markham.ca — fetchable, but 180KB of framework; needs a targeted extractor
  • canada.ca 处理时间表 — fetchable, but rewrites hundreds of numbers weekly — pure diff noise

So Vancouver, the Ontario newsroom, Edmonton and Markham are gaps: readers there are covered by federal and provincial pages only. When a workable source turns up it goes into the table above.

2. Every run is logged

Four firings a day (03:00 / 09:00 / 15:00 / 21:00 Calgary time). A run that publishes nothing still writes a row: a stopped pipeline and an idle one look identical unless something keeps a heartbeat.

When (MT)Sources readItems seenPublishedErrors
09-08 15:0084nothing—
09-08 09:00812/u/federal-government-exten/u/federal-government-exten—
09-08 03:0081nothing—
09-07 21:0080nothing—
09-07 15:0085nothing—
09-07 09:0080nothing—
09-07 03:0081nothing—
09-06 21:0080nothing—
09-06 15:0082nothing—
09-06 09:0080nothing—
09-06 03:0080nothing—
09-05 21:0082nothing—
09-05 15:0080nothing—
09-05 09:0080nothing—
09-05 03:0080nothing—
09-04 21:00830nothing—
09-04 15:0080nothing—
09-04 09:0088nothing—
09-04 03:0080nothing—
09-03 21:00819nothing—
09-03 15:0082nothing—
09-03 03:0000nothing—
09-02 15:0000nothing—
09-02 03:0000/g/student-aid-nb—

“Sources read” is 0 for every run before 3 September 2026: the watchers shipped that day, and before it the engine only wrote guides. Those zeroes are history, not a fault.

3. What came in, and what we did with it

Most government releases have nothing to do with settling in: ministers' schedules, grants to companies, ribbon cuttings. Rules drop those for free; what survives goes to a model that decides whether it is worth writing. Every verdict is kept with its reason, including the wrong ones — a filter that keeps no record of its refusals is a filter nobody can check.

not ours to write187in 30dpublished4in 30d

4. What went out

5. Who actually read it

Over 30 days the server logged 669 page opens. That is not a reader count — much of it is crawlers taking one page and leaving. Split out below.

24
sessions read 2+ pages
the closest honest proxy for people
263
views from Canada
402
single-page sessions
mostly crawlers
0
helpful votes
from readers

Most-opened pages

Where they came from

  • US385 views · 368 sessions
  • CA263 views · 44 sessions
  • HK7 views · 1 sessions
  • RU5 views · 4 sessions
  • SG3 views · 3 sessions
  • DE2 views · 2 sessions
  • CN2 views · 2 sessions
  • IE1 views · 1 sessions

On Google

2062 impressions and 5 clicks that week, average position 48.6; 937 distinct queries surfaced us over 28 days.
2026-08-30 → 2026-09-05 · Google's data lags ~2 days

Nobody has pressed the helpful buttons yet, and this line stays rather than being hidden. The site is new and has almost no audience; saying so is more honest than a flattering number.

6. What we have failed at

Probably the most useful section here. A transparency page that lists only wins is not one.

Topics we tried to source and could not

Writing anything requires fetching the official page. These are subjects whose jurisdiction's site we could not read — mostly Yukon, Nunavut, NWT, PEI and New Brunswick. They are retried on a widening interval rather than abandoned, because government sites do get rebuilt.

  • student-aid · NL — 3 attempts：抓到 1 个活页面，顺着链接又找了一层，仍然没有一页真讲这件事
  • student-aid · YT — 3 attempts：6 个候选网址全抓不到正文（死链/挡爬虫/JS 空壳），从域名首页顺链接也没找到
  • student-aid · NU — 3 attempts：6 个候选网址全抓不到正文（死链/挡爬虫/JS 空壳），从域名首页顺链接也没找到
  • childcare · NU — 3 attempts：6 个候选网址全抓不到正文（死链/挡爬虫/JS 空壳），从域名首页顺链接也没找到
  • disability-assist · BC — 3 attempts：抓到 1 个活页面，顺着链接又找了一层，仍然没有一页真讲这件事
  • disability-assist · QC — 3 attempts：6 个候选网址全抓不到正文（死链/挡爬虫/JS 空壳），从域名首页顺链接也没找到
  • disability-assist · MB — 3 attempts：6 个候选网址全抓不到正文（死链/挡爬虫/JS 空壳），从域名首页顺链接也没找到
  • disability-assist · SK — 3 attempts：6 个候选网址全抓不到正文（死链/挡爬虫/JS 空壳），从域名首页顺链接也没找到
  • disability-assist · NB — 3 attempts：6 个候选网址全抓不到正文（死链/挡爬虫/JS 空壳），从域名首页顺链接也没找到
  • disability-assist · PE — 3 attempts：6 个候选网址全抓不到正文（死链/挡爬虫/JS 空壳），从域名首页顺链接也没找到
  • disability-assist · NL — 3 attempts：抓到 1 个活页面，顺着链接又找了一层，仍然没有一页真讲这件事
  • student-aid · PE — 3 attempts：6 个候选网址全抓不到正文（死链/挡爬虫/JS 空壳），从域名首页顺链接也没找到

Known, and not hidden

  • The readership is tiny. The multi-page session count above is all of it.
  • Beyond Calgary and Toronto, most city halls cannot be fetched, and Vancouver blocks us outright. See section 1.
  • The pages are machine-written. The gates below catch what they can, not everything; the button at the foot of each guide reports a page that needs re-checking.

7. The gates every page passes

  1. Report, never advise. Immigration and tax advice are regulated professions here and nobody on this site is licensed, so pages report what the official page says and never what you should do. A sentence-level check enforces it.
  2. Quotations are verified character by character. Every change note quotes the English sentence it rests on, and that quotation is checked against the fetched page. A model can paraphrase a page it half remembers; it cannot quote one it never read.
  3. Every link is fetched before publishing. Any official URL in the text is requested for real; a dead one fails the whole page.
  4. A first read is a baseline, never an event. When a page is first watched, its current state is a baseline. In a feed, only the last 72 hours count as news.

Something wrong? tell us — or use the re-check button at the foot of any guide. Both show up in the counts above.