Source Health Dashboard
Automated link checks of every source URL cited in the Conspirafy corpus, run in small batches throughout the day. Last check 12 minutes ago.
Download source-health.json.gz for programmatic access.
7,602
Total sources
4,988
65.61%
Alive
655
Broken
1,787
Couldn't verify
172
Not yet checked
3,875
With archived copy
30-day alive % trend
Historical snapshot from daily cron runs (migration 0339).
Top broken domains
Domains with the most sources confirmed broken (page not found, or the host could not be reached, on repeated checks). Sources we could not verify are not counted here.
| # | Domain | Broken sources | View domain |
|---|---|---|---|
| 1 | bbc.com | 52 | Domain page → |
| 2 | snopes.com | 28 | Domain page → |
| 3 | justice.gov | 24 | Domain page → |
| 4 | cdc.gov | 21 | Domain page → |
| 5 | theguardian.com | 20 | Domain page → |
| 6 | search.worldcat.org | 19 | Domain page → |
| 7 | politifact.com | 18 | Domain page → |
| 8 | who.int | 15 | Domain page → |
| 9 | nsarchive.gwu.edu | 13 | Domain page → |
| 10 | fda.gov | 12 | Domain page → |
| 11 | vault.fbi.gov | 11 | Domain page → |
| 12 | adl.org | 10 | Domain page → |
| 13 | factcheck.org | 10 | Domain page → |
| 14 | bbc.co.uk | 8 | Domain page → |
| 15 | nasa.gov | 8 | Domain page → |
| 16 | badastronomy.com | 7 | Domain page → |
| 17 | bellingcat.com | 7 | Domain page → |
| 18 | gov.uk | 6 | Domain page → |
| 19 | archives.gov | 6 | Domain page → |
| 20 | propublica.org | 6 | Domain page → |
| 21 | en.wikipedia.org | 6 | Domain page → |
| 22 | archive.org | 6 | Domain page → |
| 23 | worldcat.org | 6 | Domain page → |
| 24 | nist.gov | 6 | Domain page → |
| 25 | fbi.gov | 6 | Domain page → |
| 26 | dni.gov | 5 | Domain page → |
| 27 | ico.org.uk | 5 | Domain page → |
| 28 | imf.org | 5 | Domain page → |
| 29 | nature.com | 4 | Domain page → |
| 30 | ftc.gov | 4 | Domain page → |
| 31 | cbsnews.com | 4 | Domain page → |
| 32 | oig.justice.gov | 4 | Domain page → |
| 33 | vanityfair.com | 4 | Domain page → |
| 34 | pbs.org | 4 | Domain page → |
| 35 | gao.gov | 4 | Domain page → |
| 36 | rollingstone.com | 4 | Domain page → |
| 37 | vox.com | 4 | Domain page → |
| 38 | smithsonianmag.com | 3 | Domain page → |
| 39 | politico.eu | 3 | Domain page → |
| 40 | hrw.org | 3 | Domain page → |
| 41 | fullfact.org | 3 | Domain page → |
| 42 | un.org | 3 | Domain page → |
| 43 | latimes.com | 3 | Domain page → |
| 44 | si.com | 3 | Domain page → |
| 45 | bea.aero | 3 | Domain page → |
| 46 | usgs.gov | 3 | Domain page → |
| 47 | scientificamerican.com | 3 | Domain page → |
| 48 | time.com | 3 | Domain page → |
| 49 | skepticalinquirer.org | 3 | Domain page → |
| 50 | cnn.com | 3 | Domain page → |
How alive-checking works
- A scheduled job checks a small batch of source URLs about every 15 minutes. It sends a HEAD request, retries with GET if needed, and waits up to 15 seconds. It runs at most 4 checks at once and never more than one at a time per host.
- A final 2xx or 3xx response → alive.
- Page not found (404/410), a domain that does not exist, or a refused connection, on 2 checks in a row → broken.
- Everything else is couldn't verify: access refused (401, 402, 403, 429, 451), server errors, timeouts and TLS errors. Many news and journal sites block automated checks, so we do not call these broken.
- Sources that have not been checked yet are counted separately as not yet checked.
- Where a source cannot be reached, we link to an archive.org copy if one exists. 3,875 sources currently have an archived copy.
- Paywalled pages may return 200 even when content is inaccessible. We judge by HTTP status only, not content accessibility.