Agent Ready by WorkPods

Runs every on-page SEO, GEO and AXO check on one URL: title and meta, headings, canonical, robots directives, hreflang, structured data, links, speed, and whether AI crawlers can read the page without JavaScript. Use it to check a page or to confirm a fix. With the Chrome extension connected it also compares the raw HTML with the rendered page and measures LCP and CLS; without it, those checks come back not_applicable and a note says so. Returns url (the final URL after redirects), status (its HTTP status), findings, and notes on what could not be checked. Each finding has checkId (a WorkPods check id), layer (seo, geo, axo), group, status (pass, partial, fail or not_applicable), severity (critical, high, medium or low on a fail or partial, null otherwise), a one-sentence summary and evidence (selector, url, value, note).

Crawls a site politely (robots.txt honoured, one request at a time) from the start URL, its sitemap and any pinned pages, then runs the cross-page checks: duplicate titles and content, broken links, links to redirects, orphan and deep pages, sitemap errors, soft 404s and hreflang return links. Use it for a whole-site health check; use audit_page for one page. It can take several minutes. Returns startUrl; stop (complete, limit, stopped or error); pages, each with url, status, title, wordCount and its findings; findings, one site-wide roll-up per cross-page check saying how many pages it fired on; and notes on what was left out. Each finding has checkId (a WorkPods check id), layer (seo, geo, axo), group, status (pass, partial, fail or not_applicable), severity (critical, high, medium or low on a fail or partial, null otherwise), a one-sentence summary and evidence (selector, url, value, note).

Checks the site's robots.txt, sitemaps, llms.txt and agent manifests, and gives a robots.txt verdict for each AI crawler with its role (training, search_index or user_fetch). A missing optional file such as llms.txt is not_applicable, never a failure. Returns origin; facts (robots.txt status, Content-Signal lines, aiCrawlers each with allowed true or false for the home page, the agent files found, and the Markdown version); sitemap (the files read and the URLs listed); and findings. Each finding has checkId (a WorkPods check id), layer (seo, geo, axo), group, status (pass, partial, fail or not_applicable), severity (critical, high, medium or low on a fail or partial, null otherwise), a one-sentence summary and evidence (selector, url, value, note).

Shows a page the way an AI agent reads it: the raw HTML a crawler without JavaScript gets, the rendered page text, the accessibility tree (roles and accessible names of controls) and the WebMCP tools the page registers. Use it to explain why an agent or AI answer misses content, or before a journey. Needs the Chrome extension. A null part could not be read, and an empty tool list means the page registers no WebMCP tools.

Runs every page check on the user's page and a competitor's page and lines them up check by check, with word count, headings, schema types, JavaScript dependence, AI crawler access, llms.txt and agent manifests for each. Use it to see what a rival page that outranks or out-cites the user does differently. Needs the Chrome extension. Returns one row per check with both statuses, and the facts side by side.

Reads an AI answer open in a Chrome tab (ChatGPT, Gemini, Claude, Perplexity, Google AI Mode or AI Overviews) in the user's own session. Use it after the user has asked a buyer question there; it never types or sends a prompt. Returns the engine, the prompt as shown, the brands named in order with the passage naming each, the cited URLs, and for the brand and each competitor a verdict: cited (a citation points at its domains), named, or absent. Needs the Chrome extension and an open answer tab; an error says when the tab holds no answer it can read.

Finds keyword ideas from seed terms with search volume, keyword difficulty, intent and the strength of the domains in the top 10 results, so a low difficulty score with strong domains ranking is visible. Use it for keyword research in one market. It spends from the user's own DataForSEO account inside the extension, within the monthly cap set there; a call over the cap is refused with an error saying so, and a cached answer from the last 7 days costs nothing. The agent never sees the key. Returns rows of keyword, volume, difficulty, intent and top-10 domain strength; a null figure is unknown to DataForSEO, never zero.

Lists keywords the competitors rank for in one market that the user's domain does not, or ranks far below them for. Use it to find content the site is missing. It spends from the user's own DataForSEO account inside the extension, within the monthly cap set there; a call over the cap is refused with an error saying so, and a cached answer from the last 7 days costs nothing. The agent never sees the key. Returns rows of keyword, volume, the user's position (null when it does not rank) and each competitor's position.

Summarises the backlinks of a domain or page: referring domains, domain rank, new and lost links, and top anchors. Use it to size a site's authority. It spends from the user's own DataForSEO account inside the extension, within the monthly cap set there; a call over the cap is refused with an error saying so, and a cached answer from the last 7 days costs nothing. The agent never sees the key. If the user's DataForSEO plan has no Backlinks access, an error says so.

Lists domains that link to the competitors but not to the user's domain, ranked by domain rank. Use it to find outreach targets. It spends from the user's own DataForSEO account inside the extension, within the monthly cap set there; a call over the cap is refused with an error saying so, and a cached answer from the last 7 days costs nothing. The agent never sees the key. Returns rows of the linking domain, its rank, and which competitors it links to.

Starts or looks at an agent journey: the extension opens the start URL in a scratch tab and returns what an agent can see there, the page URL, title, visible text and the controls with a ref each, plus the WebMCP tools the page registers. Start with url and task; continue with journeyId after each journey_act. Use it to test whether an AI agent can complete a key task on the site. Needs the Chrome extension. Returns journeyId, the step count (a journey ends at 15 steps) and the observation.

Takes one step in a journey started with journey_observe and returns the new observation. The extension refuses to submit a form, press a final button (pay, book, send, sign up, place an order) or sign in; a refused step comes back with stopped_by_policy and the reason. End every journey with a finish action. Needs the Chrome extension.