Every source we monitor, and every one we cannot yet

The listing is the whole archive on one 2.6 MB page (2000 → today, about 32 instruments per 30 days). It carries no ETag/Last-Modified, so we fetch it at most hourly and parse only the newest 150 rows. English and Bangla PDFs of the same circular are stored as one document.

No terms, copyright or disclaimer page is published on bb.org.bd (only a webmaster feedback link). We poll politely: at most one listing GET per hour, conditional GETs on PDFs, identified User-Agent.

robots.txt: robots.txt 302-redirects to the HTML home page: no robots rules exist.

60 circulars since 2002 (about 5 per year) in the same table shape as Bangladesh Bank. The sibling 'circular letter' page is not yet covered and PDF text layers were not sampled in research, so extraction quality is confirmed per document.

Footer shows a 'Disclaimer' label without a link; its text is unverified. Polite polling as for Bangladesh Bank.

robots.txt: robots.txt → 404 (no rules).

Coverage gap: client-rendered SPA with an internal back-office API. We use NBR's own customs SRO HTML list instead.

No terms found. The site is an Angular single-page app backed by an undocumented internal API (/api/v1/set/sro/*); we do not call private APIs without written permission.

robots.txt: unverified

Unverified: only the home page was fetched on 2026-10-12 (Apache/PHP 7.2 site); the legislation listing structure is unknown and no parser exists.

No terms found on the home page.

robots.txt: unverified

The two newest pages (80 rows) are polled; 40 SROs in the last year, 19 in the last 90 days (2026-10-12). Filenames contain Bangla letters, commas and zero-width joiners and are percent-encoded on download; PDFs were not sampled in research, so extraction quality is confirmed per document.

Terms of Use (nbr.gov.bd/terms-of-use): material is subject to copyright; linking directly to pages is allowed; loading pages into frames is not permitted. We link to the official PDF and keep a copy only as evidence.

robots.txt: robots.txt → 404 (no rules).

Unverified: the General Orders pages were seen in NBR's navigation on 2026-10-12 but not fetched (request budget). The DataTable parser probably applies; it has not been confirmed, so the source stays disabled.

Terms of Use (nbr.gov.bd/terms-of-use): material is subject to copyright; linking directly to pages is allowed; loading pages into frames is not permitted. We link to the official PDF and keep a copy only as evidence.

robots.txt: robots.txt → 404 (no rules).

Listing may lag; cross-check Gazette. Only 3 income-tax SROs appeared on the official list in the last 365 days (verified 2026-10-12), far below the normal volume, so this list is shown with a 'lagging' freshness label. 30 of 123 reference cells are in Bangla script.

Terms of Use (nbr.gov.bd/terms-of-use): material is subject to copyright; linking directly to pages is allowed; loading pages into frames is not permitted. We link to the official PDF and keep a copy only as evidence.

robots.txt: robots.txt → 404 (no rules).

The two newest pages (80 rows) are polled. Publication is bursty (budget week in June; 27 SROs in the last year, none in the last 90 days as of 2026-10-12). PDFs are Bangla Word exports with legacy-font glyph artefacts, so quotes may contain broken words and such documents are marked 'partial'.

Terms of Use (nbr.gov.bd/terms-of-use): material is subject to copyright; linking directly to pages is allowed; loading pages into frames is not permitted. We link to the official PDF and keep a copy only as evidence.

robots.txt: robots.txt → 404 (no rules).

Not fetched in the 2026-10-12 research round (request budget): vat.gov.bd, epb.gov.bd and roc.gov.bd must be verified before any coverage claim. Listed under NBR for VAT Online; EPB and RJSC belong to the Ministry of Commerce.

unverified

robots.txt: unverified

565 rows on one page (notifications, directives, orders, rules and amendments mixed); the instrument type is inferred from the subject or the filename prefix and recorded as such. Some PDFs are scanner images without a text layer: they are shown as 'scanned, summary pending' until Bangla OCR exists.

Footer 'All Rights Reserved © Copyright 2026'; the Disclaimer opens a modal whose text was not captured (unverified). We link to the official PDF and keep a copy only as evidence.

robots.txt: robots.txt: 'User-agent: *' / 'Disallow:' (allow all) — the only real robots.txt among the researched sources.

Not ingested: 41 entries, mostly internal HR office orders (retirements, pay scales) embedded as iframes — low regulatory value and no parser built.

Same as BSEC notifications.

robots.txt: robots.txt allows all.

Live since 2026-10-12 (National Portal v2 adapter, every 12 hours). The listing has no date column: a full date printed in the title is used; otherwise the date stays unknown, because the portal's record times only show the 2026-08-20 migration. Some PDFs have broken Bangla conjunct order, so in-text search can miss words.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Slice 2: authoritative and about 15 issues per week, but list titles are generic, every issue needs a second request for its part PDFs, and the PDF text layer is Bijoy/ASCII-encoded Bangla (not Unicode). Requires a Bijoy→Unicode converter and per-issue subject extraction before any coverage claim.

Footer text 'Terms of Use | Privacy Policy' has no hyperlink; 'Copyright © Bangladesh Government Press'. Terms unverified.

robots.txt: robots.txt → 404 (no rules).

Slice 2: 629 records (12 per page, POST search form); SROs are gazetted here, but the same Bijoy-encoded PDF problem applies and the page is 400 KB because of embedded ministry dropdowns.

Same as the gazette archive: no linked terms page.

robots.txt: robots.txt → 404 (no rules).

Live since 2026-10-12 through the shared National Portal v2 adapter: one request per poll for the 20 newest rows, every 12 hours. Staff/HR orders, recruitment, tenders and meeting notices are skipped on purpose. Many PDFs are scanned signed copies; those are stored, linked and flagged for review, and matching uses the title until OCR exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Not polled: the listing works with the portal adapter, but its newest entry is dated 2019-04-08 (checked 2026-10-12). Import/export policy orders are tracked through the Ministry of Commerce instead.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Live since 2026-10-12 through the shared National Portal v2 adapter: one request per poll for the 20 newest rows, every 12 hours. Staff/HR orders, recruitment, tenders and meeting notices are skipped on purpose. Many PDFs are scanned signed copies; those are stored, linked and flagged for review, and matching uses the title until OCR exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Live since 2026-10-12 through the shared National Portal v2 adapter: one request per poll for the 20 newest rows, every 12 hours. Staff/HR orders, recruitment, tenders and meeting notices are skipped on purpose. Many PDFs are scanned signed copies; those are stored, linked and flagged for review, and matching uses the title until OCR exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Live since 2026-10-12 through the shared National Portal v2 adapter: one request per poll for the 20 newest rows, every 12 hours. Staff/HR orders, recruitment, tenders and meeting notices are skipped on purpose. Many PDFs are scanned signed copies; those are stored, linked and flagged for review, and matching uses the title until OCR exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Coverage gap: the TLS handshake fails from our client (sslv3 alert handshake failure on 2026-10-12). We never downgrade TLS; the source will be retested from the production runtime and monitored via the Gazette meanwhile.

unverified (site unreachable from our client)

robots.txt: unverified

Coverage gap: Cloudflare managed challenge (HTTP 403, 'Just a moment…') for non-browser clients. We do not bypass bot challenges; BEZA would need to allow-list our User-Agent, or the source stays manual.

unverified (Cloudflare challenge blocks non-browser clients)

robots.txt: unverified

Unverified: the home page loads (Cloudflare-fronted Laravel site) but the /circular listing was not fetched on 2026-10-12, so its structure and dates are unknown.

No terms found on the home page.

robots.txt: unverified

Live since 2026-10-12 through the shared National Portal v2 adapter: one request per poll for the 20 newest rows, every 12 hours. Staff/HR orders, recruitment, tenders and meeting notices are skipped on purpose. Many PDFs are scanned signed copies; those are stored, linked and flagged for review, and matching uses the title until OCR exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Live since 2026-10-12 through the shared National Portal v2 adapter: one request per poll for the 20 newest rows, every 12 hours. Staff/HR orders, recruitment, tenders and meeting notices are skipped on purpose. Many PDFs are scanned signed copies; those are stored, linked and flagged for review, and matching uses the title until OCR exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Live since 2026-10-12 through the shared National Portal v2 adapter: one request per poll for the 20 newest rows, every 12 hours. Staff/HR orders, recruitment, tenders and meeting notices are skipped on purpose. Many PDFs are scanned signed copies; those are stored, linked and flagged for review, and matching uses the title until OCR exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Coverage gap: idra.gov.bd has no DNS record; idra.org.bd is a client-rendered Nuxt app whose content loads from webportal.idra.org.bd. Headless rendering is out of policy for v1, so insurance circulars are manual until IDRA offers a feed.

No terms found.

robots.txt: unverified

Coverage gap: MRA publishes circulars on a free-form static page, not the portal's notice table, so the shared adapter finds no rows (checked 2026-10-12). Manual until a dedicated parser exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Reference only: a database of consolidated Acts served as UTF-16 HTML with a reCAPTCHA-protected search. Used for lookup links, never polled for changes.

No copyright or terms text found; search is behind reCAPTCHA, which we do not bypass.

robots.txt: unverified

Live since 2026-10-12 through the shared National Portal v2 adapter: one request per poll for the 20 newest rows, every 12 hours. Staff/HR orders, recruitment, tenders and meeting notices are skipped on purpose. Many PDFs are scanned signed copies; those are stored, linked and flagged for review, and matching uses the title until OCR exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).

Live since 2026-10-12 through the shared National Portal v2 adapter: one request per poll for the 20 newest rows, every 12 hours. Staff/HR orders, recruitment, tenders and meeting notices are skipped on purpose. Many PDFs are scanned signed copies; those are stored, linked and flagged for review, and matching uses the title until OCR exists.

National Information Portal (v2) site: no terms/copyright page linked; footer disclaimer only (Bangla). Polite polling with an identified UA.

robots.txt: robots.txt returns a 404 HTML error page (no rules).