Search found 668 personal blogs on github.io — a census, not a complete registry
Asked:
Find every personal blog on github.io
A search sweep as of 1 September 2026: 21 diverse queries returned 3,327 unique result pages, semantic classification retained 900 pages that look like personal blogs, and those consolidate to 668 distinct github.io hosts. A person name was extracted for 375 hosts and a blog name for 111.
Why “every” is impossible. github.io has no authoritative public index of personal blogs; search engines cover only what they have crawled and chosen to rank, many sites are unindexed or unlinked, and the collection changes continuously as pages appear and disappear. Names, blog titles, and topic descriptions below are machine-extracted from search results and may be imperfect or missing; unresolved cells are left blank.
Only 27% of retained result pages survived consolidation as distinct hosts — most blogs surfaced on several queries, and the busiest single host accounts for 31 matched pages: canro91.github.io
A person name could be extracted for 375 of 668 hosts (56%); a distinct blog title for only 111 (17%) — most github.io blogs present under the owner's name alone: ericmjl.github.io
Topic descriptions were recovered for 396 hosts (59%), ranging from research notebooks to Minecraft diaries: 11011110.github.io
545 of 668 hosts appeared on exactly one matched result page — a hint that many more single-appearance blogs sit just below the search horizon: github.io
Directory of all 668 discovered hosts
Search the table by host, person, blog name, or topic. Identities and descriptions are search-derived and potentially imperfect.
Host
Person
Blog name
Topic (search-derived)
Pages
Method: web search census of github.io personal blogs, snapshot 2026-09-01. 21 queries → 3,327 unique result pages → 900 pages classified as personal blogs → 668 distinct hosts (one table row each). “Pages” counts matched search-result pages per host. Person names (375), blog names (111), and topic descriptions are machine-extracted and unverified; blanks mean no confident extraction. Summary totals come from precomputed metrics, not recounted. Not exhaustive: unindexed and newly created sites are absent by nature of search discovery.