Frequently Asked Questions
Plain-English answers about AEO, AI visibility, Dynamic SEO from Google Calendar, Auto-Publish to WordPress and Webflow, our MCP server, and pricing.
AgenticSEO is an AI visibility platform and execution layer. It audits a website, runs your buyer phrases against ChatGPT and Google Gemini to see whether you or a competitor gets named, then writes the title tags, meta descriptions, JSON-LD schema, llms.txt, and FAQ answers that close each gap — and publishes them to WordPress, Webflow, HubSpot, or a Git repository once you approve.
AgenticSEO is built and supported by Traction Gap Partners, LLC, led by founder Bruce Cleveland, author of Traversing the Traction Gap and Market Engineering: Because Markets Don't Build Themselves.
Yes. The product is named AgenticSEO and lives at agenticseo.live. "TGP Agentic SEO" was an earlier internal name for the same application from Traction Gap Partners.
Yes, two ways, neither of which needs a credit card. The free check audits three pages of your website with no sign-up and hands back publishable assets immediately. A 14-day trial then opens the full product — tracked prompts, the audit-to-publish workflow, and auto-publish on eligible tiers. Nothing is charged unless you choose a plan.
AgenticSEO is an Answer Engine Optimization (AEO) platform from Traction Gap Partners. It measures how ChatGPT and Google Gemini currently describe and cite your brand, then generates the metadata, Schema.org JSON-LD, llms.txt, and FAQ content that closes the gaps so those assistants recommend you when buyers ask.
Run the free AI visibility check: enter your domain and AgenticSEO asks ChatGPT and Google Gemini the questions your buyers ask. You get back whether your brand was named, which competitor was named instead, and whether AI crawlers can read your site at all. No card required.
Finished work, not just a score. Along with the visibility result, the free check produces three ready-to-publish assets from your own pages — an answer-ready FAQ block with matching FAQPage JSON-LD, a schema and AI-tag snippet for your homepage, and one answer-ready page draft for a buyer question your site doesn't answer — plus a dated 30-day fix/publish calendar built from your findings. Everything is shown in-page with copy buttons, emailed to you, and downloadable as a Word document.
Yes. The assets are yours to paste into your site tonight — no license, no attribution, no card. A paid plan is for scale and automation: every page audited instead of a few, fixes written for the whole site, and publishing done for you on a schedule.
Agencies need three things a tracker rarely has: client-scoped websites, white-label reporting, and an API. AgenticSEO provides all three — a separate, isolated website per client, branded .docx reports under the agency's own logo and colors, and both an MCP server and a REST API so audits and fixes run headlessly across the whole roster.
Yes. AgenticSEO is an execution layer, not just a dashboard. After the audit it writes the title tags, meta descriptions, JSON-LD schema, llms.txt, and FAQ passages, publishes them to WordPress, Webflow, HubSpot, or a Git repository behind an approval gate, then re-measures on the next scheduled run.
Traditional SEO tools optimize for Google's ten blue links. AgenticSEO optimizes for AI assistants that answer without sending a click — measuring citation rate across ChatGPT and Google Gemini, and producing a full six-layer discoverability stack per site (Classical SEO, Schema.org JSON-LD, AI-crawler access, answer-readiness, llms.txt, and sitemap.xml). Few platforms do all of this; we bundle it into one loop.
B2B software companies, agencies, and category creators who need to be cited by AI assistants — not just ranked in Google. Founders use the $29 one-time report; growth teams and agencies use the subscription tiers for scheduled audits, Dynamic SEO, and multi-client management.
Yes — agentic search optimization is another name for agentic SEO. It emphasizes the behaviour being optimized for: agentic search, where an AI assistant runs the queries, reads the pages, and synthesizes one answer for the user instead of returning a list of links. The deliverables are identical — entity-clear JSON-LD, verified AI-crawler access, self-contained answer passages, an llms.txt index, and measured citation share on real buyer prompts.
Agentic AI for SEO (or agentic AI SEO) is the technology underneath the practice: AI agents that act rather than suggest. An agent crawls your site, measures how assistants answer your buyers' questions, ranks the gaps by impact, drafts the metadata, schema, llms.txt, and FAQ content each one needs, publishes on approval, and verifies the result on the next scheduled run.
A branded .docx report — rendered under your own (or your agency's / OEM partner's) logo and colors — covering the six-layer discoverability audit (Classical SEO, JSON-LD, AI-crawler access, answer-readiness, llms.txt, and sitemap.xml), an AI-visibility snapshot across ChatGPT and Google Gemini, a prioritized content-gap list, and draftable FAQ content for the gaps. Subscribers also get scheduled re-audits and Dynamic SEO auto-publish.
The Traction Gap Framework (TGF) is Bruce Cleveland's operating model for scaling category-defining companies. Every AI prompt we author, competitive comparison we run, and page recommendation we make is framed through TGF market-engineering language — so the metadata reads like a serious growth thesis, not generic SEO copy.
AEO is the practice of structuring your site so AI answer engines — ChatGPT and Google Gemini — cite you when they respond to buyer prompts. It replaces keyword-density thinking with entity, schema, and llms.txt-driven optimization.
We run your tracked prompts against real assistants (ChatGPT and Google Gemini, via our AI gateway), then parse each response for brand mentions, URL citations, and competitor references. Results roll up into a citation-rate score you can trend over time on the Visibility dashboard.
ChatGPT (OpenAI) and Google Gemini run on every plan — together they cover the majority of buyer-facing AI queries and represent the two most divergent retrieval systems, so wins on both tend to predict wins on the rest. Claude, Perplexity, and Google AI Overviews are included in the Enterprise tier, and can be enabled for Pro accounts on request to TGP; those engines carry an additional per-run usage cost.
Yes. Claude, Perplexity, and Google AI Overviews are part of the Enterprise tier. Pro accounts can have them enabled on request to TGP — they carry an additional per-run usage cost because each run consumes third-party engine capacity. ChatGPT and Google Gemini remain standard on every plan at no extra cost. Contact us and we'll enable the additional engines for your account.
A content gap is a tracked prompt with no strong matching page on your site — meaning AI assistants have nothing solid of yours to quote. We list every gap on the Gaps tab and can draft ready-to-paste FAQ Q&A pairs (via generate-faq-doc) so you can publish an answer under /faq in minutes.
We read the website you (or your OEM partner) provide and build a compact snapshot of what the site actually publishes — page topics, product names, and any brand context on file (industry, audience, value proposition, focus keywords, known competitors). That snapshot is sent to our AI, which returns a de-duplicated set of buyer-intent prompts a real ChatGPT or Gemini user would type — mixed across awareness ("what is X"), consideration ("best X for Y", "X vs Y"), and decision ("pricing for X", "is X SOC2 compliant") stages. Prompts are anchored to what the site actually sells; no generic keyword bait. If only the website is available, that alone is enough to seed a first prompt set — quality improves as the brief and competitors are filled in.
Yes. The generator reads the company brief (industry, audience, value proposition, focus keywords, known competitors) and the tracked-competitor list, so filling those in narrows and sharpens the output. You can also pass a free-text hint at generation time ("focus on procurement personas", "skip pricing prompts"). Every generated prompt lands in a review queue where you accept, edit, or dismiss it before it becomes a tracked prompt.
llms.txt is an emerging plain-text index intended to point AI systems and agents toward the pages you consider important. Adoption by the major crawlers is not confirmed, so treat it as cheap insurance rather than a ranking lever — it works alongside classical SEO, Schema.org JSON-LD, AI-crawler access, and visible answer-ready content. AgenticSEO generates a site-wide llms.txt tuned to your positioning as one layer of that stack.
We produce a 6-layer discoverability stack plus one optional annex. Layer 1 Classical SEO (title, description, canonical, OpenGraph). Layer 2 Schema.org JSON-LD (Organization, Product, FAQPage, DefinedTerm, etc.). Layer 3 AI-crawler access (robots.txt allow-list and fetch checks for GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended, Bingbot, Applebot-Extended, CCBot). Layer 4 answer-readiness (whether each page directly answers the buyer questions AI assistants ask — direct-answer paragraph, FAQ block, byline + date, heading hierarchy, external citations). Layer 5 site-wide llms.txt at /llms.txt. Layer 6 sitemap.xml with normalized URLs and lastmod dates. Annex A (optional, off by default): experimental ai:* meta annotations — self-documenting audit trail only, not a published standard and not consumed by mainstream LLMs.
AI-crawler access checks whether your robots.txt, meta robots, and server actually let the AI assistants fetch your pages — GPTBot (ChatGPT), Google-Extended (Gemini), PerplexityBot, ClaudeBot, and friends. If those user-agents are blocked or your CDN 403s them, none of your other SEO work reaches the LLMs. We surface this as its own layer because it is the single most common silent reason a well-optimized site still gets zero citations.
Answer-readiness scores whether each page directly answers the buyer questions AI assistants are asking — clear question-shaped headings, short quotable answers near the top, entity coverage, and inline citations. It is the bridge between your tracked prompts (Content Gaps) and the on-page fixes AgenticSEO writes, so a page that looks fine to Google but is unquotable by an LLM gets flagged before it costs you a citation.
No — not in any documented way as of 2026. There is no published web standard for an ai:* meta-tag namespace, and none of the major AI search surfaces (Google AI Overviews, ChatGPT / OpenAI, Anthropic Claude, Perplexity) publish support for consuming these tags at query time. Google explicitly ignores unknown <meta name> values for ranking. Modern LLMs extract meaning from the raw page text via retrieval-augmented generation, not from bespoke metadata. AgenticSEO ships an ai:* emitter as Annex A of the report — off by default — for tenants who want a machine-readable audit trail of what our analysis believed each page was about. The four levers that actually move AI citations are Classical SEO, Schema.org JSON-LD, AI-crawler access in robots.txt, and answer-readiness in the visible copy.
Yes — every page recommendation and this /faq page itself emit FAQPage JSON-LD. FAQPage schema is a useful structured format for clear, direct-answer content, which is the shape assistants tend to quote.
Enterprise AI-visibility platforms are measurement systems — they tell you where your brand is missing from ChatGPT and Google Gemini. AgenticSEO does the same tracking and goes further: it generates the metadata, Schema.org JSON-LD, llms.txt, and FAQ drafts that close each gap, with approvals and safe publishing. Pro is $99/mo for multi-engine tracking versus $399/mo at comparable enterprise tiers. See /vs/ai-visibility-platforms for the full comparison.
AgenticSEO already covers both. It tracks citations and share of voice across ChatGPT and Google Gemini, and pulls Google Search Console for CTR, ranking, and cannibalization signals — feeding both streams into the same execution queue. Many teams can avoid a second subscription by consolidating on AgenticSEO.
AgenticSEO Starter includes ChatGPT + Google Gemini citation tracking, GSC integration, and site audits at a lower price point than enterprise AI-visibility platforms, which typically charge more for ChatGPT-only tracking and add-on engines. See /pricing for current tier pricing.
Most options fall into two groups: classical SEO suites that audit titles, links, and rankings, and AI-visibility products that measure how assistants describe you. AgenticSEO sits in the second group but adds the write step — it produces the metadata, JSON-LD, and FAQ blocks, and can publish them to your CMS after review.
Measurement-first platforms report where assistants skip you. AgenticSEO runs the same tracking and then generates and publishes the fix — schema, llms.txt, metadata, and answer-ready page drafts — through an approval queue. See /compare/profound-alternative for the capability-by-capability table.
Not entirely. Keep a classical suite for backlinks and large-scale keyword research; AgenticSEO covers answer-engine visibility, structured data, crawler access, and publishing. Many teams run one of each.
Run a free 3-page preview, then connect your CMS (WordPress, Webflow, GitHub, or HubSpot) with your own credentials. From there each audit produces a queue of proposed changes you approve before anything is published.
Buyers increasingly shortlist through assistants before they visit a website. If ChatGPT or Gemini names a competitor when asked your category question, you lose the evaluation before it starts — which is what the tracked-prompt report measures.
llms.txt is an emerging plain-text index at /llms.txt that points AI systems toward the pages you consider important. Adoption by major crawlers is unconfirmed, so treat it as cheap insurance layered on top of crawler access, schema, and answer-ready copy. AgenticSEO generates and maintains the file for you.
robots.txt controls what a crawler is allowed to fetch; llms.txt is a curated content map. robots.txt is a real standard with enforcement — llms.txt is a convention, which is why AgenticSEO audits crawler access first and treats llms.txt as one later layer.
It shouldn't. Classical metadata and schema still do the heavy lifting; llms.txt is a low-cost addition, not a substitute. AgenticSEO publishes all of these layers together so the choice never comes up.
In the domain root, at /llms.txt. AgenticSEO checks that the file is present, well-formed, and refreshed as your pages change.
Yes. The audit generates a site-wide llms.txt from your actual crawl, and — if you have connected a CMS or repository — can publish it after you approve it. You can also use the standalone generator at /tools/llms-txt-generator.
The current field includes AgenticSEO, Profound, Peec AI, Otterly.AI, AthenaHQ, Writesonic Goldie, HubSpot's AI Search Grader, and Semrush's AI toolkit. The buyer's guide at /compare/best-aeo-tools compares them row by row with vendor sources.
Classical suites optimize for click-through from search results. AgenticSEO optimizes for being the quoted source inside an answer — crawler access, structured data, and short quotable passages — and can ship those changes to the page.
Yes. AgenticSEO runs from a single site up to multi-client agency and OEM portfolios, with per-client websites, page caps by plan, and an API/MCP surface on custom plans.
You see which buyer prompts name you versus a competitor, why (blocked crawlers, missing schema, unquotable copy), and you get the fixes drafted and published rather than a report to hand to engineering.
Less than it used to. With a CMS connected, approved changes publish directly; without one, you get copy-ready snippets and deployment instructions. Larger structural changes still benefit from a developer.
Look for one that reads your real pages instead of asking you to fill in a form. AgenticSEO derives Organization, SoftwareApplication, Article, and FAQPage markup from the crawl, then publishes it after review. The standalone FAQ schema generator is at /tools/faq-schema-generator.
Copy-paste generators stop at the snippet. AgenticSEO detects what markup a page is missing, drafts the JSON-LD in context, and can push it live — so the fix does not sit in a ticket queue.
Organization and SoftwareApplication (or Product) for entity clarity, FAQPage for direct-answer content, and Article with a byline and date for guides. AgenticSEO generates these and flags pages where the type is wrong for the content.
Yes. The audit re-derives markup for each page, so invalid or stale JSON-LD is replaced rather than patched, and you approve the change before it publishes.
Yes, within your plan's monthly page volume. Larger sites usually run scheduled audits by section so structured data stays consistent across thousands of pages.
At pre-seed, the practical starting point is a real audit rather than a retainer: the free 3-page preview shows whether assistants can reach and quote you, and the $29 one-time report covers a full site.
A classical suite tells you where you rank; AgenticSEO tells you whether ChatGPT and Gemini can cite you and fixes what blocks them. For a small site with no backlink profile yet, the second question usually matters sooner.
Sometimes — assistants reward clear, well-structured, specific answers, so a small site can win narrow buyer prompts even without authority. Broad category prompts still favor incumbents.
No. Connecting a CMS is a credentials step, and approved fixes publish from the app. Teams without a CMS connection use the copy-ready output instead.
The loop closes. Instead of a dashboard that reports the same gap month after month, each tracked gap has a drafted fix, an approval, a publish, and a re-check.
Peec AI, Otterly.AI, AthenaHQ, Writesonic Goldie, Semrush's AI toolkit, and AgenticSEO. The difference with AgenticSEO is that it publishes fixes as well as tracking citations — see /compare/profound-alternative.
Semrush is a broad SEO and SEM suite; tracking-first AEO products report AI mentions. AgenticSEO measures answer-engine visibility and then updates the site — schema, metadata, llms.txt, and answer-ready drafts.
It drafts every fix automatically and publishes on approval; nothing goes live unreviewed. That publish step is what most measurement-first tools leave to you.
No — it is complementary. Keep your classical suite for backlinks and keyword research, and use AgenticSEO for AEO execution.
Categories where buyers research through assistants before talking to sales — B2B software, professional services, and agencies managing many client sites — since those are the queries where being uncited is expensive.
Make sure the assistants can read you and answer the category question correctly, then track it. AgenticSEO reports which prompts name you versus competitors, which is a defensible number to show in a deck.
Moving from a working product to repeatable traction depends on market awareness. If assistants recommend competitors on your core buyer questions, awareness stalls — so tracked-prompt coverage is a useful early indicator.
Semrush reports search performance and Profound reports AI mentions; AgenticSEO reports AI mentions alongside the remediation history, so you can show both the gap and what you did about it.
Share of AI voice on category prompts, whether the brand is named or replaced by a competitor, and the trend over time. AgenticSEO benchmarks all three against your tracked competitor list.
It can fix the source: when an assistant describes you with outdated information, the app flags the prompt and drafts the on-page correction. No product controls an assistant's output directly.
Free for a 3-page preview, $29 for a one-time full-site report, and $39/mo (Starter) for continuous tracking with 25 tracked prompts. Current tiers are on /pricing.
It is self-serve and priced per plan rather than quoted, and it includes the publish step — which matters when there is no engineer free to implement a report's recommendations.
Semrush has an AI toolkit for visibility reporting, but it does not draft and publish the on-page fixes. If you already pay for Semrush, keep it and add AEO execution.
Tracked prompt count, pages audited per month, how often you re-run, and whether you use auto-publish. Plans are sized on those dimensions rather than seats.
No setup fee. Connecting a CMS or repository takes credentials you already have, and the free preview needs nothing but a URL.
They give an outside-in read: how often assistants name the company on category prompts, and who they name instead. That is harder to overstate than self-reported marketing metrics.
Citation frequency on buyer-intent prompts, whether the answer names a competitor instead, sentiment, and the trend line. AgenticSEO tracks these per prompt and per competitor.
Search rankings lag; assistant answers reflect what buyers see today. AgenticSEO reports the assistant side, which classical suites only partly cover.
Yes — it surfaces cases where assistants describe a company inaccurately, attribute its features to a competitor, or cannot find it at all.
Fix crawler access first, then add the missing schema and a direct-answer passage on the pages that should own the prompt. AgenticSEO drafts those changes and re-checks the prompt after publishing.
Compare AgenticSEO, Profound, Peec AI, Otterly.AI, AthenaHQ, Writesonic Goldie, and Semrush's AI toolkit. The matrix at /compare/best-aeo-tools lists capabilities and prices with links to each vendor's own pages and the date they were checked.
Semrush is stronger on keyword and backlink data; AgenticSEO is built for answer-engine visibility and ships the on-page fixes. They overlap less than they look.
It drafts the fixes automatically and publishes them to your CMS once approved, then re-checks the affected prompts on the next run.
It becomes hard to do by hand at scale: thousands of pages and hundreds of prompts need scheduled audits, consistent structured data, and per-brand reporting, which is what the Agency and Enterprise plans cover.
SEO aims to win a click from a results page; AEO aims to be the source quoted inside a generated answer. The technical work overlaps — crawlability, structure, clarity — but AEO adds crawler access for AI agents and short quotable passages.
AI startups can measure share of AI voice by tracking how often their brand is cited or recommended across Large Language Models like ChatGPT and Google Gemini. AgenticSEO provides specific visibility metrics that identify which prompts trigger brand mentions and where competitors currently hold the majority of the AI response space.
Traditional SEO tools like Semrush focus on keyword rankings and search engine result page positions. In contrast, AgenticSEO measures AI visibility by analyzing the generative output of AI assistants to ensure your brand is part of the synthesized answer provided to the user.
While Profound and Ayzeo offer insights into AI engine performance, AgenticSEO is designed to both identify visibility gaps and fix them by publishing optimized content directly to your website. This integration allows startups to move quickly from measurement to execution within a single platform.
Yes, measuring and improving AI visibility is critical for reaching Minimum Viable Repeatability (MVR) in a market where buyers increasingly use AI assistants for discovery. AgenticSEO helps startups engineer their market presence to ensure they are recognized as a category leader by generative engines.
Startups should prioritize the citation rate, the sentiment of the AI's description, and the accuracy of the product features mentioned. AgenticSEO tracks these data points to provide a comprehensive view of how AI models perceive and recommend your brand compared to the rest of the market.
To improve search ranking in AI models, SaaS companies should identify specific prompts where their brand is currently missing from AI-generated answers. AgenticSEO automates this process by identifying visibility gaps and generating structured FAQ content that directly addresses the queries AI assistants prioritize. This content must be factually dense and formatted for easy ingestion by Large Language Models.
While traditional tools like Semrush focus on keyword volume and blue-link search engine results pages, AgenticSEO focuses on AI visibility and Answer Engine Optimization (AEO). Our platform not only identifies the questions buyers are asking AI models but also automates the publishing of those answers directly to your website to secure citations. This moves beyond keyword tracking into active AI influence.
Yes, the AgenticSEO framework is designed to scale from SMBs to Enterprise organizations. For SMBs, it provides a way to compete with larger brands by capturing niche AI queries, while Enterprise teams use it to maintain market share across thousands of complex product prompts. The goal is to ensure the brand is cited as a primary source regardless of company size.
AI assistants prioritize content that is structured, authoritative, and directly answers the user's intent without excessive marketing fluff. By using AgenticSEO to publish clear Q&A pairs that mirror actual user prompts, SaaS brands increase the likelihood that models like ChatGPT and Google Gemini will pull their data into the generated response. This creates a direct link between the AI's answer and the brand's website.
AgenticSEO is built to both identify and remediate visibility gaps. Unlike competitors like Profound or Ayzeo that may only provide analytics, our application can fix identified issues by publishing optimized content directly to your website. This reduces the time between discovering a visibility gap and appearing in AI search results.
Dynamic SEO turns your Google Calendar into a publishing pipeline: create a calendar event with a topic and target URL, and AgenticSEO generates a fully optimized page (metadata, JSON-LD, agentic tags, body copy) and auto-publishes it to WordPress or Webflow at the event's start time. Most measurement-first AEO tools do not make this the core workflow.
Bring-Your-Own-CMS. You connect WordPress, Webflow, a GitHub-backed static site, or HubSpot once; from then on, any recommendation AgenticSEO makes can be pushed live with one click — no copy-paste, no engineering ticket. HubSpot can be connected on every plan (Free through Agency). Pushing artifacts back into any CMS is a Pro/Agency benefit.
WordPress (via application password), Webflow (via API token), GitHub-backed static sites (via a repo PR), and HubSpot (via a Private App access token). The Connections page lists each integration; more CMS targets are on the roadmap.
A scheduler publishes what you already wrote. Dynamic SEO generates the page from a calendar entry — full metadata, JSON-LD, agentic tags, body draft, and internal linking — using TGF-framed prompts and your live brand context, so the page ships AEO-ready without a human writer in the loop.
No. You paste a WordPress application password, a Webflow API token, a GitHub fine-grained PAT, or a HubSpot Private App access token into Connections once, then everything happens from the app UI.
Only if you explicitly select an existing URL as the target. By default, new recommendations create new URLs (typically under /faq, /resources, or /solutions), and each push shows you a diff-style preview before it goes live.
Yes. On Pro and Agency, every connected site can be put on a re-analysis schedule (weekly, monthly, quarterly). Each run refreshes visibility scores, detects new gaps, and can auto-open Dynamic SEO drafts for the gaps that scored highest.
Yes. The free preview runs a real 3-page AgenticSEO audit on your own site with no account and no card required. Sign up (still free) to run up to 3 previews per company before you decide whether to buy a $29 one-time report or subscribe.
A full audit of your site — up to 25 pages by default — with all three metadata layers, a site-wide llms.txt, a competitive AI-visibility snapshot, and a downloadable .docx branded to your (or your agency's) company. No subscription, no recurring charge.
Scheduled re-audits, AI-visibility trend tracking across ChatGPT and Google Gemini, the Gaps and Benchmarks tabs, Auto-Publish (Pro+), Dynamic SEO from Google Calendar (Pro+), agency multi-client management (Agency), and the Finn AI assistant embedded in the app.
Typically 3–8 minutes for a 25-page site. We batch page fetches (10 client-side, 5 server-side in parallel) so a full crawl finishes well inside a coffee break instead of the multi-hour queues most SEO tools use.
Yes. The Agency tier gives you a client switcher, per-client briefs, isolated visibility runs, and billing that stays on your Agency account. Each client sees only their own data.
Yes. Every visibility run parses assistant responses for competitor mentions alongside your own, so you can see who is being cited when you're not. Competitors are auto-discovered on your first analyze run — you can then add missing ones or hide any that aren't real competitors from the Prompts tab.
No. On your first analyze run, Finn auto-discovers competitors from your brand context (industry, positioning, product names) and tags them as source='auto'. You can add manual competitors alongside them, and both are used in future runs. MCP/REST integrators can trigger the same discovery headlessly via the discover_competitors tool.
Yes. Every competitor chip in the Competitors card has an × button that soft-deletes the entry so it's excluded from future runs and Share-of-Voice math. Removed entries can be re-added manually at any time, and re-running Auto-discover will not resurrect ones you've explicitly hidden.
Yes. Every report exports as a .docx branded to your (or your agency's / OEM partner's) company — logo, primary and accent colors, and 'Prepared by' label. Subscribers can also export company data as a GDPR-compliant zip from Manage → Data, and integrator API access is documented at /docs/api.
You decide. Set a default retention window on Dashboard → Manage → Report retention that applies to every new report, and override individual reports from the Reports list using the Extend menu. Tier caps: Starter up to 6 months, Pro up to 12 months, Agency up to 24 months, Enterprise unlimited. Each report shows its exact expiry date; expired reports display an Expired screen with a one-click Extend so you can restore access before running a fresh audit.
Benchmark trend analysis (share-of-AI-voice, citation coverage, prompt win rate over time) is only as long as your oldest kept report. Extending retention on the reports you care about — a quarterly baseline, a pre-launch snapshot, or a competitor comparison — lets Benchmarks draw a longer trendline instead of resetting when older reports expire.
Yes. AgenticSEO exposes a production MCP (Model Context Protocol) server so agents like Claude, Cursor, and custom LLM clients can drive audits, read reports, and trigger publishes over token-based auth. See /docs/api for the REST parity endpoints.
Yes. Connect Google Search Console in Connections and the Benchmarks tab adds a Google Search panel: clicks, impressions, click-through rate and impression-weighted average position for the window, change chips vs. the previous period, and your top queries and pages ranked by impressions so page-two near-misses surface. These are measured Google numbers, not third-party estimates. It is included on every plan, costs no credits, and is also available over the MCP server and REST API (get_gsc_performance / GET /v1/gsc/performance).
Yes. Every core action — start audit, fetch report, list gaps, push to CMS — is available at /functions/v1/api with a project token. Full reference lives at /docs/api.
It upgrades the Google half of your report from measured to reported. You get real clicks, impressions, click-through rate and impression-weighted average position straight from Google instead of estimates; page-two near-misses surface (terms already earning impressions but almost no clicks), which are usually the highest-leverage fixes; and it is free, read-only and included on every plan, so it costs no credits and we never change anything in the Google account. Live search checks still work without it — they show where a term ranks, but they cannot show impressions, clicks or CTR. Running both is the most complete picture.
That screen comes from the Google app that starts the connection, not from Search Console. By default connections run through our shared Google app, and Google only lets accounts we have already approved sign in through it. Open the "Your Google OAuth app" card on the Coverage page and register your own Google Cloud OAuth client (Web application type, plus the authorized redirect URI we display and the read-only Search Console scope on your consent screen). After that your own consent screen decides who may connect. Agencies register one app and every client they manage inherits it; a client can still register its own, which takes precedence. Client secrets are stored encrypted and never shown again, and existing connections keep working on the app that created them.
Yes. Every crawl runs against a named market, so a UK, German, or Japanese site is fetched from that country with a matching Accept-Language header, and Google SERP and AI Overview captures use that market's language and country parameters. In the app the market is inferred automatically from the domain (example.co.uk → United Kingdom, example.de → Germany) and falls back to the United States. Over the MCP server and REST API you can name it explicitly with the country parameter (ISO alpha-2, e.g. "gb") on discover_site / POST /v1/discovery and analyze_site / POST /v1/reports.
41 markets: Argentina, Australia, Austria, Belgium, Brazil, Canada, Chile, China, Colombia, Croatia, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, India, Indonesia, Ireland, Israel, Italy, Japan, South Korea, Latvia, Lithuania, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Romania, Saudi Arabia, Singapore, Slovenia, Spain, Sweden, Switzerland, Thailand, the United Kingdom, and the United States. If an integration asks for a market outside that list the API returns an error rather than quietly crawling from the US, so a report never claims to be local when it isn't.
In Connections, click Connect WordPress, enter your site URL, and paste an application password (Users → Profile → Application Passwords in WP admin). We store the credential encrypted and only use it for pushes you explicitly approve.
In Connections, click Connect Webflow, paste a site API token, and pick which collection new pages should be written to. Publishes create a draft item first so you can preview before going live.
HubSpot is available on every plan. In HubSpot go to Settings → Integrations → Private Apps → Create a private app (name it "TGP AgenticSEO"), grant the "content" scope (required for updating Site Pages, Blog Posts, and uploading llms.txt), copy the Access Token (starts with pat-), and paste it into the HubSpot card on Connections. Revoke it any time from Private Apps — TGP holds no HubSpot OAuth app.
HubSpot AEO tracks visibility across ChatGPT, Gemini, and Perplexity and returns advisory recommendations ("create this blog post, update that page") that you still have to write and publish. TGP tracks ChatGPT and Google Gemini on every plan — plus Claude, Perplexity, and AI Overviews with Enterprise or on request to TGP — with prompts grounded in your actual site crawl — then executes the fix: crawler-access audit, JSON-LD, answer-readiness engineering, ai:* tags, llms.txt generation, and one-click push into your HubSpot Site Pages and Blog Posts. HubSpot tells you the gap; we close it.
All CMS credentials are stored encrypted at rest, scoped per-user via Row-Level Security, and never leave our backend. Crawl data is stored under your company only. Delete Account (Manage → Data) is a full self-serve GDPR wipe with ownership transfer and payment detach.
We can't render sites that require a signed-in session (LinkedIn, Facebook, Instagram, X/Twitter, TikTok, Quora, Glassdoor, Nextdoor), video/audio shells (YouTube, Vimeo, Spotify), or infinite-catalog marketplaces with aggressive bot mitigation (Amazon, eBay, Etsy, Walmart, Airbnb, Booking, Expedia, Zillow, Redfin, Yelp, TripAdvisor). There's no TOS-compliant workaround — public crawlers see the login wall, not the content. The UI shows a warning before you submit and the MCP/REST API returns HTTP 422 with code `unsupported_target`. Analyze the brand's own marketing site, docs, or newsroom instead.
We use the Lovable AI Gateway to route to Google Gemini (primary generator) and OpenAI gpt-5-mini for live visibility measurement across ChatGPT and Gemini. You do not need to bring your own API keys.
One-time site report is $29 (no subscription). Subscriptions include Starter (tracked prompts, copy-ready outputs), Pro (more prompts + auto-publish write-back to WordPress, Webflow, GitHub, and HubSpot + Dynamic SEO), and Agency (highest prompt cap, multiple client seats included, per-seat overage). Prompt Pack add-ons are available on Pro and Agency. Current tier pricing lives at /pricing.
Starter includes 25 tracked prompts per month, Pro includes 100, and Agency includes 300. If you outgrow your tier's cap, add a Prompt Pack of +100 prompts for $19/mo on Pro or Agency — no need to jump tiers. Enterprise plans are custom.
Pro and Agency include one-click auto-publish to WordPress, Webflow, GitHub, and HubSpot. Starter is copy/paste only — you'll get the same JSON-LD, llms.txt, and FAQ output, but you deploy it yourself. Auto-publish is a direct-plan feature and is not exposed through the OEM/MCP integration surface.
Our promise is an outcome, not a refund window: monthly plans carry the first-month delivery guarantee — if we don't ship the agreed number of published items in your first month, your next month is free. You also see real output before paying, because the free check hands you three publishable assets. We do not advertise a blanket 14-day refund, but you can request a one-time discretionary refund via support if a report didn't run correctly.
On a monthly plan we commit to publishing a set number of items to your site in your first month — the target is shown on every eligible pricing card and on your dashboard. If we come in short and nothing was blocked on your side, your account is flagged automatically at period end and TGP applies the following month as a credit.
One approved fix pushed live (title, meta description, JSON-LD, llms.txt), one published answer-ready FAQ, or one page published to your site. Each one is written to a delivery ledger with the page URL and a timestamp when it goes live, so the count is auditable rather than a marketing claim. Drafts, downloads, and copy/paste output you deploy yourself are not counted.
Starter, Pro, and Agency subscriptions, monthly and annual. It does not apply to the free check, the 14-day trial, or the one-time site report, which are one-off deliverables rather than an ongoing publishing commitment. Enterprise and OEM agreements carry their own terms.
The clock pauses rather than paying out. We can only publish what you approve to a target we can reach, so if no publish target is connected or nothing has been approved, the guarantee period is marked as waiting on you — your dashboard card names which of the two is blocking so it's never a surprise.
Yes. From Manage → Plan & Billing you can upgrade, downgrade, pause, or cancel at any time. Tier changes are prorated through Stripe.
The free preview is unlimited-in-value for evaluation: 3 previews × 3 pages per company at no cost, no card. Ongoing use requires the $29 one-time report or a subscription.
Yes — two ways in, both without a credit card. The free check audits three pages of your website and hands you publishable assets immediately, with no sign-up. Separately, a 14-day trial opens the full product — tracked prompts, the audit-to-publish workflow, and auto-publish on eligible tiers — and no card is required to start it. Nothing is charged unless you choose a plan when the trial ends.
No. Neither the free three-page check nor the 14-day trial asks for a card. You only enter payment details when you buy the $29 one-time report or start a subscription.
Yes at the entry point, and the comparison is not like-for-like. Starter tracks buyer prompts across ChatGPT and Google Gemini, audits your pages, and connects Google Search Console for less than enterprise AI-visibility platforms typically charge for ChatGPT-only tracking with add-on engines. AgenticSEO also writes and publishes the fix — metadata, JSON-LD, llms.txt, and answer-ready FAQs — where monitoring-only platforms stop at the report and you still pay someone to implement it. Current tier pricing is at /pricing and the capability-by-capability matrix is at /compare/best-aeo-tools.
Run a real 3-page audit on your own site in under two minutes — no card required — or browse the full glossary of AgenticSEO terms.
Last updated:
The guides, comparisons, and free tools behind AgenticSEO.