Shows which AI crawlers may fetch a page
robots.txt and Content-Signal, bot by bot, and how each page splits into the passages assistants quote.

GLLmolfar crawls your site on your own computer, runs hundreds of checks and puts what to fix first. Windows 10+ and macOS on Apple silicon.


01
GLLmolfar reads robots.txt and the sitemaps, and results fill in while it crawls.

02
Hundreds of checks, each with why it matters, how to fix it and where to look.

03
The fix list puts what costs the most first. Export to CSV, Excel, JSON or XML.

robots.txt and Content-Signal, bot by bot, and how each page splits into the passages assistants quote.

See what a bot reads without a browser, and what only appears once the page's scripts run.

Topics by meaning with EmbeddingGemma on your computer, translations kept apart, similar pages side by side.

A linking map by section shows where internal weight goes and where links are missing.

Clicks and positions next to the pages that earn them. Traffic diagnosis finds the days clicks changed, splits the change into impressions, position and the rest, marks Google updates and forecasts seven weeks, formulas shown, no AI model. Crawls run on a schedule, and history never reads a smaller crawl as lost pages.

Crawls, pages and embeddings stay on your computer. Nothing about the sites you crawl is sent to us.
A local MCP server: your agent or any MCP client starts crawls and reads the findings. Model calls are counted first.
Content studio drafts from your facts with LM Studio, Ollama or the built-in model, and flags invented numbers.

Find the pages closest to a topic before a new draft repeats one you already have.

Experimental, with the free GLLmolfar Connector: before and after, and every change can be rolled back.
Audit before a migration
Crawl the old site, keep the export, crawl again after the move.
Check a core update
Traffic diagnosis marks Google updates and splits the change in clicks.
Open pages to AI assistants
Find blocked AI bots and text that shows only after JavaScript.
Settle competing pages
See which pages split one query, which to keep and how to merge.
Crawl client sites
An agency plans crawls per client and exports to Excel or Sheets.
Check a draft first
A writer sees whether a draft overlaps a page that already exists.
Fix WordPress titles
Titles, descriptions and canonicals in one batch, each one undoable.
Hand the audit to an agent
Your agent crawls through MCP while you watch each step.

The short answers. The docs go into the detail.

Free to start.* Bigger sites go on paid levels.
No. Crawls, pages and embeddings stay on your computer. An anonymous usage ping (version, system, language, country) can be switched off in Settings.
No. The crawl and its checks need no model, and embeddings run on your computer. A model is optional, for drafting and planning.
The builds are not code-signed yet, so Windows and macOS ask before they open them. The download page says what to do.
On the GetLoopLoop download page, through a personal link.
* We are still researching what people need from the app, so levels and limits will change.