About this project

Adult FYI takes thin RSS posts (typically one or two sentences), finds related sources via web search, synthesizes a fact-grounded article with a local language model, and stages it for human approval before publication.

Run the pipeline (CLI):

php /home/much/public_html/adultfyi.2much.net/cli/run.php

Pulls the newest unprocessed item from https://adultfyi.com/feed/. First-time setup requires the schema:

mysql -u much_easy_ga_stats -p'...' much_easy_ga_stats < /home/much/public_html/adultfyi.2much.net/schema.sql

Review queue

Pending review — synthesized articles awaiting approval
Approved articles
Old vs. new — side-by-side comparison of every story

Pipeline stages

Ingest — fetch + parse RSS, dedupe by URL hash
Search — Brave query built from the title (filters out source domain + social sites)
Fetch + extract — pull each result, strip to article text (rejects pages under 400 chars)
Synthesize — four AI calls: body, meta, FAQ, entities
Review — admin UI shows side-by-side; human approves or rejects

Key files

config.php — API keys, DB, model choice, admin password
lib/synth.php — synthesis prompts and JSON-LD builders
cli/run.php — orchestrates one full pipeline run
admin/ — HTTP basic auth + side-by-side review UI
article.php — public render with full SEO + JSON-LD (Article + FAQPage)