Skip to main content
Glama

Server Details

Check a landing page on a real phone and desktop, get the fixes, re-check after you publish.

Ownership verified
Status
Healthy
Uptime
100.0% over the last 24 hours
OAuth
Works in Glama
Last Tested
Transport
Streamable HTTP · MCP 2025-06-18
URL
ChatGPT
–
Claude
LandingBoost

TDQS

A3.9/5.0

Scored across 13 tools

Disambiguation4/5

Most tools target clear, distinct objects: get_actions (actionable items), get_report (full report), get_scan (job status), get_sites/list_scans/get_watches (listings), and the write-side scan_site/watch_site/mark_done/recheck_page are all unambiguous. The main overlap is get_sites versus get_watches, since get_sites already reports a 'watched' flag, and get_actions versus get_report, though the description explicitly steers agents to get_actions for the large report. get_next_action and get_actions could be confused at a glance, but the descriptions draw the line clearly.

Naming Consistency4/5

All names are snake_case and verb-first, with a coherent verb vocabulary (get_, list_, scan_, watch_, mark_, recheck_, write_, suggest_). The only deviation is list_scans versus get_sites/get_watches, which are also collection getters but use 'get' instead of 'list'. Readable and predictable overall.

Tool Count5/5

Thirteen tools is well within the well-scoped range and each maps to a real step in the workflow: page suggestion, scanning, polling, reading actions, marking done, rechecking, watching, and AI retry. No tool looks like filler or a trivial wrapper.

Completeness4/5

The core lifecycle is covered end to end: suggest_pages → scan_site → get_scan → get_actions → mark_done → recheck_page, plus watch_site/get_watches and write_wording for the failed-AI case. Gaps are minor but real: there is no unwatch/remove-watch tool, no way to cancel a running scan, and no delete/manage for stored sites.

Publisher details

Operator
LandingBoost · Publisher source
Vendor relationship
First-party · Publisher source
Privacy policy
Not provided
Trust center
Not available
Restrictions
Needs a LandingBoost account (free sign-up). Free plan: 1 scan a month with the full report. Re-checks after a fix and daily watching need Pro ($19/month), Studio ($49/month) or Lifetime ($149 once). · Publisher source

Available Tools

13 tools
get_actionsWhat to changeA
Read-only
Inspect

What to change. teardown: what the report leads with, the 3 to 5 fixes found on the whole homepage and the pricing, on a phone and on desktop, each with its id (td:…, for mark_done), its status, what was found, the fix, the new copy and agent_text, plus apply_all (one message for all of them); on Free the first fix is open in full and the others stay locked until the owner unlocks the report ($9) or has a paid plan. checked: every area that was checked and its result (passing areas included). do_first: problems measured on the page and the wording that fixes one, with the exact current and new text and their evidence (often empty: nothing measured is urgent). smaller: the rest of the report, worth doing but not urgent: new wording backed by AI answers or competitors, checks worth a look, then polish. also_fix: failed search and crawler settings, safe to apply. ask_owner_first: measured layout, contrast or slow-page problems; propose these and wait for the owner. Every item has agentText, a one-change instruction. format=text returns the same brief as plain text.

Inputs: scan_id (required), format.

get_next_actionThe one thing to do nextA
Read-only
Inspect

The single most important thing to do next across all your sites: a watched site that broke, then the first open fix the report leads with (teardown), then a must-fix still open, then new wording still open, then an item marked done that recheck_page can confirm, then anything else open. Returns the site, the item with its agentText, why it was picked, and progress on that report.

Inputs: no named parameters.

get_reportFull reportA
Read-only
Inspect

The full report of one of your scans: the teardown, actions plus every measured check, pages and limits. Large; prefer get_actions.

Inputs: scan_id (required).

get_scanScan statusA
Read-only
Inspect

Status and progress of one of your scans, and whether its AI part (new wording) came back (ai_part). When status is complete or partial, call get_actions.

Inputs: scan_id (required).

get_sitesYour sitesA
Read-only
Inspect

Every site you scanned, as on the dashboard: the latest report (scan_id, date, whether the new wording came back), how many "Do these first" items are done or confirmed, the next item on that report, and whether the site is watched (open changes).

Inputs: no named parameters.

get_watchesWatched sites and their changesA
Read-only
Inspect

Your watched sites: status (ok, changed, break), open changes with before and after, the last checks, and days without a break.

Inputs: no named parameters.

list_scansYour scansA
Read-only
Inspect

Your recent scans (scan_id, site, status, date), newest first.

Inputs: no named parameters.

mark_doneMark an item doneAInspect

Mark one item of get_actions done (done=false reopens it), as the owner does on the report page. Use it when the owner says an item is done, or for items recheck_page cannot confirm. Teardown fixes (id td:…) and items listed under do_first or smaller are accepted; ask_owner_first and also_fix items, and unknown ids, return an Unknown item_id error. A later recheck_page result overrides the mark.

Inputs: item_id (required), scan_id (required), done.

recheck_pageCheck a page after publishingAInspect

After the owner publishes a change: load one page of a scan again (phone and desktop) and run the same checks. Each item on that page comes back fixed, improved, still_open or visual_check (changed, compare by eye); get_actions and the report page show the result and tick fixed items off. Pro, Lifetime, or a report opened with its $9 unlock (otherwise it returns an error that says so): no scan from the limit and no AI. About 30 seconds; one at a time, up to 5 page checks per scan a day. Speed, site-wide items (AI search, facts across pages) and the teardown fixes (td:…) are not re-measured here: use mark_done, or a new scan.

Inputs: scan_id (required), page_url (required).

scan_siteScan a websiteAInspect

Start a LandingBoost scan of a public website: phone and desktop screenshots of up to 3 pages (2 on Free; homepage first; call suggest_pages to choose them, or homepage and pricing are picked automatically; the pricing page is always scanned when the site has one), measured mobile, copy and search checks, and new wording for the problems found. Returns a scan_id at once; the scan takes about 3 to 5 minutes. Poll get_scan every 30 to 60 seconds, then call get_actions. Counts toward the scan limit: use it for a new site or after big changes (a new page or layout). After a single fix, call recheck_page instead (free).

Inputs: url (required), pages.

suggest_pagesSuggest pages to scanA
Read-only
Inspect

Pages worth scanning on a website, from its sitemap and homepage links: homepage, pricing, features, comparison and other first-level pages, each with its role and where it was found. recommended=true marks the default pick (homepage and pricing). Free and does not count toward the scan limit. Pass the chosen URLs to scan_site as pages.

Inputs: url (required).

watch_siteWatch a scanned siteAInspect

Start watching the site of one of your scans: every day its main pages (the scanned ones plus product pages, up to 10) are re-read and compared with today's state. Changes are reported as break (a page stopped answering, gained noindex, lost its heading, or prices now disagree between pages), drift (title, description, heading or prices changed) or note. Free; no scan is used. As many sites as the plan watches: Free 1, Pro as many as the sites it covers, Lifetime 5.

Inputs: scan_id (required).

write_wordingWrite the missing new wordingAInspect

When get_scan or get_actions shows ai_part.status "failed" (our AI provider was busy or refused the request), write the buyer questions and new wording again from the pages the scan already saved. No new capture and no scan from the limit. Takes about a minute: poll get_scan until ai_part.status is "complete", then call get_actions. Only for a report whose AI part failed: on a report whose AI part is already complete it returns an error ("already complete") and changes nothing.

Inputs: scan_id (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      B
      maintenance
      Enables deterministic delivery QA for AI-generated web pages by opening URLs in Chromium, checking mobile viewport, text-scaling, and safe-area behavior, probing primary CTAs, and returning screenshots plus machine-readable evidence.
      1
      215 PyPI
      1
      MIT
    • A
      license
      Not graded
      quality
      Not graded
      maintenance
      Runs live, URL-synced mobile and desktop sessions of a web or responsive UI and returns measured findings — contrast ratios, undersized touch targets, overflow, theme drift — as text with the offending selector attached, so an agent fixes the exact element instead of guessing from a screenshot. A shared panel lets a human pin a note and frame that the agent reads back, keeping both in front of the same screen.
      95 npm
      MIT
    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables checking a public website before launch and returning a scored pass, warn, or fail report with a fix for every problem found.
      MIT
    Try in Browser

    Glama MCP Gateway

    Add one secure layer between your agents and this server.

    Resources