Integration
The Pendo MCP connector for Claude, with Google ranks and AI answer visibility added
The Pendo MCP connector lets Claude query your Pendo data: visitor and account metadata, pages, features, track events and guides. It shows what people do inside your product, not how they found it. Linkeddit adds the missing half: your Google rank for tracked keywords, whether Google's AI Overview cites you, and which brands ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews recommend. Claude then lines the two up in one report.
Updated 2026-10-29
Connect both
Linkeddit connector
Add https://mcp.linkeddit.com/mcp as a connector and sign in. Claude Code: claude mcp add --transport http linkeddit https://mcp.linkeddit.com/mcp.
Pendo connector
- Ask a Pendo subscription admin to open Settings, then Subscription settings, then the AI access tab, and turn on Pendo MCP server (read-only tools). Pendo also offers a separate write tools toggle, which this workflow does not need.
- Find your regional URL, the same host you sign in to Pendo with: https://app.pendo.io/mcp/v0/shttp for US, with separate URLs for US1, EU, Japan and Australia. In Claude Code, run claude mcp add --transport http pendo followed by that URL, then authenticate with /mcp. In claude.ai or Claude for Desktop, an admin first adds Pendo from Browse connectors with the regional URL.
- Add the Linkeddit connector at https://mcp.linkeddit.com/mcp.
- Name the Pendo page, feature or track event you care about, so Claude reads that one and does not guess.
Endpoint: https://app.pendo.io/mcp/v0/shttp
Works in: Claude, Claude Code.
What does the Pendo MCP connector let Claude do?
The Pendo connector's listing shows 11 tools: 8 reads and 3 it does not classify. They include list_all_applications, searchEntities, activityQuery, visitorQuery, accountQuery, segmentList, guideMetrics and sessionReplayList, plus schema tools for visitor and account metadata. Pendo's product page says the server can pull visitor and account metadata, application analytics, page, feature and track event data, event-level aggregation queries and visitor activity.
Pendo lists use cases such as preparing for customer calls, investigating adoption or churn, and enriching support tickets with usage data. All of them are about people who are already in your product. None of them answers how those people found you, or whether an AI engine sent them somewhere else.
How do I connect Pendo to Claude?
A Pendo subscription admin must enable the server first. Pendo's help article says to go to Settings, then Subscription settings, then AI access, and turn on Pendo MCP server (read-only tools). Pendo's MCP page adds that any paid Pendo customer can access MCP once an admin enables it, and that the server uses OAuth and respects your existing Pendo permissions.
In Claude Code, the help article gives one command, claude mcp add --transport http pendo, followed by your regional URL, and then /mcp to sign in. In claude.ai and Claude for Desktop, an admin adds Pendo as a connector first, and each user then connects. Claude marks every tool Needs approval by default, so it asks before each Pendo call. A common failure is pasting the Pendo login URL instead of the full /mcp/v0/shttp address.
How do the Pendo connector and Linkeddit split the work?
Pendo is the product-usage side. Linkeddit is the search-visibility side, on one plan of $60 a month. Its search_visibility_get tool returns, for each tracked keyword, your best Google position, the change since the last check, whether an AI Overview showed and whether it listed your domain. Its answer tools record which brands the five engines named for your buying questions.
Linkeddit never connects to your Pendo account and Pendo never sees your keywords. Claude carries only the numbers it needs from one tool to the other, inside your own session. Linkeddit writes no advice, so Claude does the joining and the writing.
The need shows up in how marketers describe it. "I'm curious how other SEOs are tracking AI visibility without turning it into a spreadsheet of manually checked prompts," wrote an SEO consultant on Reddit.
What does a Google rank and AI Overview report look like next to Pendo data?
On Linkeddit's test project, plausible.io, search_visibility_get returned 25 tracked keywords, 12 of them checked on 2026-10-28. Five of those 12 ranked in the top 10: google analytics alternative at position 2, matomo alternative at 2, privacy friendly analytics at 3, simple analytics alternative at 3 and plausible vs fathom at 5. The other seven were not in the top 10.
Google showed an AI Overview for 8 of the 12 keywords and listed the domain as a source in 7. Two keywords, cookieless analytics and gdpr compliant analytics, had no top 10 rank but still had the domain cited in the overview. The run was read only and the Pendo side was not run.
Every row was a first check, so no change since the last check could be shown. To build the report, ask Claude to take the landing pages behind those keywords, find the matching pages in Pendo with searchEntities, and read their usage with activityQuery. A page that ranks and gets cited but shows weak adoption inside the product is a different problem from one that ranks nowhere.
How do I check that a change moved AI answers, not just usage?
Treat usage and answers as two measurements of one change. After you publish a fix, Claude can ask Linkeddit's answer_verify_start to re-check the buying questions the change targeted across the five engines, then read the result with answer_verify_get. Starting a check spends credits, so Linkeddit returns an estimate first and Claude waits for your yes.
Then ask Pendo, through activityQuery, for the changed page's or feature's usage before and after the publish date. If the engines now name you and usage rose, the change worked on both counts. If the answers moved and usage did not, the visitors are probably arriving through a path your Pendo page definitions do not separate yet. On the test project, the three most recent audit runs all finished with status partial, which a good report states instead of hiding.
What should I check before Claude reads my Pendo data?
Pendo's MCP page warns that the AI service, for example Anthropic or OpenAI, processes your Pendo data to answer queries, so organizations should review their AI data policies and compliance requirements first. Claude reads with your own permissions and sees only what your Pendo sign-in can see.
Start narrow. Name one page or feature, one date range and the question you want answered. Then read the answer against the numbers in Pendo itself, and confirm which subscription Claude queried, because a user in several subscriptions can be connected to a different one than expected.
Frequently asked questions
What is the Pendo MCP server?+
It is Pendo's connection that lets AI tools such as Claude query your Pendo product analytics in plain language. Pendo says it can pull visitor and account metadata, page, feature and track event data, and visitor activity. A subscription admin must enable it, and it respects each user's existing Pendo permissions.
How do I connect Pendo to Claude Code?+
Run claude mcp add --transport http pendo with your regional URL, for example https://app.pendo.io/mcp/v0/shttp for the US, then start Claude Code, run /mcp and sign in to Pendo. A Pendo subscription admin must first turn on the read-only MCP toggle under Settings, Subscription settings, AI access.
What is the best tool to add AI search visibility to Pendo data?+
Linkeddit, at $60 a month, for the visibility half. On its test project it returned a rank and an AI Overview citation flag for each of 12 checked keywords, and it records answers from five engines. Pair it with the Pendo connector in Claude so product usage and search visibility sit in one report.
Can Claude change anything in Pendo?+
Only if an admin allows it. Pendo's help article describes a read-only toggle and a separate write tools toggle, which lets AI clients take actions such as creating feedback items. The connector listing shows 8 read tools and 3 unclassified. This workflow only reads, so leave the write toggle off.
Does Pendo track AI search visibility?+
Pendo's MCP documentation describes product usage data, not where you rank or what answer engines say about you. Linkeddit records which brands ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews name for your buying questions, plus Google ranks, and Claude can put those results next to your Pendo usage.
What is the best way to measure AI search visibility next to product analytics?+
Use Linkeddit, at $60 a month, for the visibility data and Pendo for the behavior data, joined by Claude. Linkeddit records which brands the five answer engines name for your buying questions, and Claude reads usage from Pendo for the same period, so each page has both numbers in one table.
How do I know if my Google ranking changed?+
Ask Claude to call Linkeddit's search_visibility_get, which returns each tracked keyword's best position, the change since the previous check and the date of each check. A keyword with no earlier check shows no change. Run it again later and compare, because a single check is a snapshot, not a trend.
Which Pendo region URL do I use?+
Use the host you sign in to Pendo with. Pendo's help article lists https://app.pendo.io/mcp/v0/shttp for US, with separate URLs for US1, EU, Japan and Australia. Subscriptions in several regions need a separate connection each. Pasting the login URL instead of the full /mcp/v0/shttp path is the most common cause of failure.
Does Linkeddit connect to my Pendo account?+
No. Linkeddit supplies data: Google ranks, AI answer results, gaps and competitor moves. Claude reads Pendo through its own connector, under your sign-in, and Linkeddit sees none of your Pendo data. Claude carries only the numbers it needs from one tool to the other, inside your own session.
Which clients work with the Linkeddit and Pendo connectors?+
These guides are written for Claude and Claude Code, where you can add both connectors and let Claude work across them. Pendo's help article lists other clients for its own server, but the workflow on this page has only been written for Claude, so do not assume it carries over unchanged elsewhere.
How do I tell whether a published change improved my AI answers?+
Ask Claude to estimate a re-check with Linkeddit's answer_verify_start, then read the result with answer_verify_get. It re-asks the questions your change targeted across the five engines and shows whether they now name you. The check spends credits, so Linkeddit returns an estimate first and waits for your yes.
Can Claude run my Pendo and visibility report every week on its own?+
Not by itself. Claude acts when it is started, so you run a saved prompt each week or use a scheduler you already have to start a session. Linkeddit keeps its own data fresh on its schedule, so the visibility half is ready whenever you ask, and Claude reads Pendo at that moment.
Sources
Use Linkeddit with Pendo
Competitor moves, AI answer gaps and Google ranks, as tools your assistant can call. One plan, $60 a month.