Blog

The source list and the chat fought over one overlay. Two panels with different z-index plus render-time visibility ended the fight.

Follow-up questions wiped the previous answer. One ThreadEntry array with functional updates turned the single-slot dialog into a real thread.

EventSource can't POST. The blog's Ask AI answers run through a plain fetch, a hand-rolled SSE parser, and one CRLF trap worth remembering.

One tiny field on the list page nearly cost a query per article. The fix is one batch IN query: two queries per page, not 1 + N.

An optional locale filter froze live search on the default locale. The culprit: bindparams validation that fires at construction, not execution.

Tag-matched related articles kept missing the point. Swap in cosine similarity, add a TLDR box, and make search live, all fail-safe against an old API.

Replace LIKE with a MariaDB FULLTEXT index, then add LLM query expansion: two layers that keep working when the index or the LLM fails.

The semantic related layer for 80 articles turned out to be a Python dict and a cosine loop, no numpy and no vector database.

Wiring env vars in Compose is only half the job; the container still needs a deliberate network path before the LLM search feature can talk to its router.