All pages

Project settings: niche, competitors, destinations and indexing

The Project group in Settings works on the project you have open; switch projects and these screens follow.

General

Rename Project takes a Project name of up to 100 characters.

Niche Details is the market definition research works from: Name, Industry, Language, Country, Target Customer, Core Problem and Related Terms. The pencil opens the form and Save Changes stores it. Your agent edits the same fields with update_niche.

Posting cadence asks for Articles per week, 1 to 14: how many you actually publish. The report's 90-day content sequence is sized to it. Your agent sets it with set_posting_cadence and reads it from get_project.

Re-map website re-fetches your sitemap and refreshes the page list, titles and descriptions on Pages. It spends no credits and runs once a day per site on Free, once an hour on a paid plan. Your agent calls it as remap_asset.

Delete Project, under Danger Zone, asks you to type the project's name, then deletes the project and all its data for good.

Excluded terms and the asset profile are not on this screen; your agent sets them with set_excluded_terms, update_asset_profile and resync_asset_profile.

Competitors

Two tables, Important and Ignored. Add opens one field, Company or domain: pick a company VarynForge already knows or type a new domain. The bin icon returns a company to neutral.

Important gives a 1.25x boost to opportunity scores for clusters where that competitor ranks in the top 10. Scores are then rescaled 0 to 100, so other opportunities can drop in relative order without losing anything. Ignored companies leave competitor analysis, opportunity scoring and the fiercest-competitors widget. New projects start with noisy domains such as YouTube, Wikipedia and Reddit already ignored.

Your agent uses list_competitors, set_competitor_importance and add_competitor_by_domain.

Destinations

Destinations are where your content lives or will live: Website, YouTube channel, Instagram, TikTok, X (Twitter), LinkedIn and Local business. Each card shows its type, a badge (Yours, Found on your site, Planned, Open slot), editable Name and URL, and what it makes and how results are measured. One found on your site carries Add to project; the bin icon removes one.

Add a destination takes a type and a name: a domain for a website, a handle otherwise. For a website or a YouTube channel, Check availability looks the name up first. It's live — it's mine records one that exists; Not built yet — plan it holds the name for later. A project holds up to 10.

Each destination shapes the briefs and drafts made for it and is the format label lint_draft files drafts under. The primary website cannot be removed or have its URL cleared: briefs, research and IndexNow anchor on it. Your agent reads them with list_destinations and adds with add_destination; editing and removal happen here. See Five channels.

IndexNow

This screen is in one of three states:

  • Not set up. Generate a key shows the file to host, https://<your host>/<key>.txt, and the exact text it holds. Host it and press Verify & enable. Generate a new key replaces it. Already hosting a key? Paste it or its file's URL into the field and press Use & verify; the file must be on the project's own site.
  • Active. Shows your host and the last URL IndexNow acknowledged.
  • Platform auto. Set with My platform already does this, for Cloudflare Crawler Hints, Yoast or Rank Math. VarynForge sends nothing, so nothing is submitted twice. Setup flags Yoast or Rank Math when it spots them.

Remove integration asks twice, then clears the key and ping history. Pings go out from mark_article_published; Publishing and attestation explains when.

Search Console

Connect Google Search Console asks Google for a grant separate from your VarynForge login: read-only Search Console access (webmasters.readonly), kept so syncs run without you. Then pick the property that covers this project's site; a likely match is marked Suggested.

Once a week VarynForge reads clicks, impressions, click-through rate, average position and the top 10 queries for each article you published with a live URL, matched by page URL. The window ends three days back, since Google's data settles late; the first sync covers 28 days. The numbers show on each article's Passport and on Earnings, and an article that earns clicks gets a lift in its priority score. Nothing scrapes search results or polls rankings.

Disconnect asks twice, revokes the grant with Google and removes the connection. Outcomes already synced stay.

Markdown version of this page, for your agent.