Each page reads paid_starts_at, free_tier, and team_plan from the CSV at render time with a 12-hour cache. When Perplexity raises a tier from $20 to $25, you edit one cell and the chip on the detail page, the cell in the matrix, and any related context referencing the price all update on the next refresh.

Yes. The roundup at /ai-research-assistants/ reads the same 15-row dataset through a list block and renders the sortable matrix. The detail page reads one matching row. Edits to the CSV propagate to both views on the same refresh, so the roundup never drifts from its children.

Update the integration_list column for that row in the CSV. The next 12-hour refresh updates the chip row on the detail page, the cell in the listing matrix, and any related context that mentions the integration. No code change, no redeploy, no roundup rewrite is required.

Each row stores its own seo_title and meta_description. The template wraps those fields into the head, so Google sees a unique title and description for every page rather than 15 near-duplicates. You can hand-write each description or let an editorial pass fill them in later.

Yes. Append a row to the source CSV with a unique slug, a complete set of column values, and status set to ready. Within 12 hours SleekRank publishes the new URL, includes the row in the listing matrix, and starts linking to it from related blocks on the other 15 detail pages.

Set the row status to archived. The URL stops appearing in sitemap.xml and the listing matrix, and the template returns a 410 Gone status with a short note plus links to two alternatives in the same research_focus tag. Google drops the URL rather than leaving a soft 404 in the index.

Each row tags itself with a research_focus value like life-sciences, general-academia, open-web. The related block reads that tag, filters the 15-row dataset to matching rows, and picks up to six entries sorted by a stable hash of slug pairs. The grid is consistent on one URL but varies between URLs cleanly.

Yes. Each row carries optional hero_variant_a and hero_variant_b fields. The template reads which variant is active from a top-level config flag or a staging query parameter. The underlying field values stay shared, only the editorial wrapper around them changes between variants in the test.