This experiment exploring how LLMs index and search "potential word forms"
If LLMs meet potential word forms while they crawling the Internet what they'll do?
Frontend: Wiki.js 2.5.311 (Node.js 24.12.0)
Backend: Wiki.js + PostgreSQL 17.7
Hosting:
Old: Railway.app
New: VPS
Monitoring: Google Search Console, Yandex.Webmaster
Version Control: Git
First part: Alice AI index a word without using Google Search Console, Yandex.Webmaster;
Google and LLMs partly index it or only can find in the Internet
after site verification with <meta> tags.
/home - home page for the project with main information
/dictionary - all potential word forms group by alphabet
/dictionary/word - word change on word form that you need.
This page include definition of this word, linguistic characteristics, examples of using etc.
{
"DB_TYPE": "postgres",
"DATABASE_URL": "${{Postgres.DATABASE_URL}}",
"PGDATA": "${{Postgres.PGDATA}}",
"PGDATABASE": "${{Postgres.PGDATABASE}}",
"PGHOST": "${{Postgres.PGHOST}}",
"PGPORT": "${{Postgres.PGPORT}}",
"PGUSER": "${{Postgres.PGUSER}}",
"POSTGRES_DB": "${{Postgres.POSTGRES_DB}}",
"POSTGRES_PASSWORD": "${{Postgres.POSTGRES_PASSWORD}}",
"POSTGRES_USER": "${{Postgres.POSTGRES_USER}}"
}
<meta name="google-site-verification" content="YOUR_GOOGLE_KEY" />
<meta name="yandex-verification" content="YOUR_YANDEX_KEY" />
Site:
Old: https://sgamga-wiki.up.railway.app/en/home