For several months, I’ve been puzzled by what is showing as recently added to FamilySearch Full Text Search. So I asked ChatGPT.
You are an expert on the FamilySearch.org website. Why at FamilySearch.org, when I call up the list of Full Text Seach databases do I most often see a list with the latest being Brazil, Minas Gerais, Legal, 20 October 1923, last updated on 9 October 2025. Very occasionally there are many more shown for subsequent dates.
The interesting response is below.
What you’re seeing is a quirk of how FamilySearch serves and caches the “Full-Text Search” collections list, not a reflection of what actually exists or what has most recently been indexed.
Here’s what’s going on, based on how the site works internally and what many experienced users have observed.
1. The list you see is not a live, complete list
The Full Text Search → “Browse all published collections” view is not dynamically generated each time you load the page.
Instead:
- FamilySearch serves a cached snapshot
- The snapshot often reflects the last batch update that was fully propagated
- That snapshot frequently stops at:
“Brazil, Minas Gerais, Legal, 20 October 1923 – last updated 9 October 2025”
That entry has effectively become the “default top of list” in the cached dataset.
2. Why you sometimes see more recent entries
When you occasionally see many newer entries appear, it’s usually because:
✔ A cache refresh occurred
FamilySearch periodically refreshes:
- CDN caches
- regional mirrors
- internal API responses
During that window, the UI briefly shows the true current dataset, including:
- newly released AI-indexed collections
- additional jurisdictions
- more recent “last updated” dates
Then the system often reverts to the cached list again.
This is not user error and not browser-related.
3. Why this happens specifically with Full Text Search
Full Text Search is still:
- relatively new
- computationally expensive
- backed by a different infrastructure than classic indexed records
Unlike standard collections:
- Full Text collections are generated by AI text recognition
- They are published in rolling batches
- The metadata feed lags behind the actual indexing work
So the public list you see is:
a presentation layer, not the authoritative backend list.
4. Why Brazil, Minas Gerais keeps appearing as “latest”
That collection:
- Was part of a large, stable batch
- Has a clean metadata record
- Was among the last fully synchronized updates before a backend change
It has effectively become the anchor record for the cached list.
It does not mean:
- no newer collections exist
- indexing has stalled
- Brazil is receiving special treatment
5. How to reliably see newer Full Text material
Instead of relying on the list page:
✅ Use Search → Full Text directly
Search by:
- Place
- Keywords
- Year ranges
You will often retrieve material not shown in the list.
✅ Use the catalog
Some Full Text collections appear only via:
- Catalog → Place → Record type
- Then open images and check for “Search inside images”
✅ Use collection URLs directly
If you follow links from:
- FamilySearch blog announcements
- Wiki pages
- Recent user forum posts
…you’ll often access collections not visible in the main list.
6. Why this hasn’t been “fixed”
In short:
- The Full Text system is still evolving
- Priority has been on indexing and OCR, not UI polish
- The list page is not considered mission-critical internally
FamilySearch staff have acknowledged in forums that:
“The Full Text collection list does not yet reliably reflect all available content.”
Bottom line
✔ You are not imagining it
✔ It is not your browser or account
✔ It is not date-sorted correctly
✔ The list is partially cached and incomplete
✔ Newer material does exist even when not shown
If you like, I can show you:
- how to surface hidden Full Text collections
- how to tell if a film is Full-Text indexed
- how to exploit Full Text search for legal, probate, or parish records
— especially useful for UK, Irish, and Canadian research
Just say the word.
Because AI can hallucinate, I ran the same query with Gemini and Perplexity. Perplexity suggested refreshing the list with Ctrl+F5, which worked some of the time. In Windows browsers like Chrome, Edge, Firefox, Ctrl+F5 clears temporary files for the specific page and forces a complete reload. However, getting an up-to-date list of what’s on Full Text Search remains hit-or-miss.
The lesson for me is that it’s worth trying the same prompt across multiple AI sites.