How to get LinkedIn post analytics, lead counts and booked meetings in Claude or ChatGPT

TL;DR: Yes. Connect the traxy MCP to Claude or ChatGPT and you can ask how your LinkedIn posts and leads are doing: it reads aggregate performance, post analytics, qualified leads per post, the content feed and meetings attributed to your leads, so the AI can put together a weekly summary or a client report in the chat. The numbers come from your traxy workspace, not from guesses, and the same connection runs the outreach to those leads: LinkedIn and email campaigns from your own connected sender accounts, with a yes in chat before anything sends.

This guide covers how it works step by step, which MCP tools read which numbers, prompts for a weekly summary and for finding the posts and tracked profiles that brought in qualified leads, how an agency builds a client report from chat, and what the numbers mean and don't. Tool names come from the traxy MCP tool catalog, and the reporting rules from the Insights and Meetings and outcomes docs, checked November 2026.

Can I ask Claude how my LinkedIn posts and leads are doing?

Yes. With the traxy MCP connected, Claude or ChatGPT reads your workspace's performance metrics, the posts of the profiles you track, the qualified leads each post brought in and the meetings matched to your leads, and answers in plain English.

That is useful because the question "how are my posts doing?" usually has two answers. LinkedIn's own analytics tells you how many people reacted or commented. traxy tells you who those people were, which of them fit your ideal customer profile (ICP), and whether any of them later booked a meeting. The MCP brings the second answer into the chat you already use.

traxy runs your entire LinkedIn GTM end to end, from finding high intent leads from social intent to enriching and researching each one, to running relevant, personalized LinkedIn and email outreach to them on autopilot. Autopilot starts with your approval: you test a campaign on sample leads and approve it once, then it runs, or you approve every message. Reporting sits across both halves:

• Finding leads. traxy watches social intent on LinkedIn: what people post, comment and like, engagers with the people, posts and company pages you track, and your own posts. A person only counts as a lead with a 50%+ match to your ICP, and every lead keeps the post or comment behind it. Post analytics and leads per post report on this half.
• Outreach. Campaigns research each lead, group leads by signal and prepare personalized LinkedIn and email sequences that send from your connected accounts. Campaign performance and attributed meetings report on this half.

How does it work, step by step?

You connect once, grant the read permissions, then ask questions. The AI calls the traxy tools behind the scenes and turns the results into an answer.

1. Connect traxy to your AI chat. In traxy, open Integrations, pick your client (Claude, ChatGPT, Cursor and others), add https://api.traxy.ai/api/v1/mcp, sign in, pick the workspaces and approve. No API key is needed. The AI client docs have the steps for each client.
2. Grant the reporting scopes. Performance, post analytics, qualified leads per post and meeting counts need analytics:read. The attributed meetings list needs leads:read, and the content feed needs content:read. Your workspace role and product access still apply, so a tool in the catalog doesn't mean it works in every workspace.
3. Track the sources you want to report on. Post analytics covers tracked profiles, so track your own LinkedIn profile, your teammates, or each client's profile. You can add them under Targeting > People in the app, or ask the AI to track them from chat.
4. Connect Google Calendar for meetings. Meeting attribution matches calendar meetings to traxy leads. Connect Google Calendar under Integrations and select the calendars to use. Without it there are no attributed meetings to read.
5. Ask a question with a scope. Name the workspace, the period and what you want back. The AI picks the tools, reads the results and writes the summary.
6. Check the answer against the source. Ask the AI to name which tool each number came from. For anything that looks off, open the same view in the app (Insights, the tracked profile, or Outbound > Meetings) with the same date range.

The reporting tools are reads. Asking how your posts are doing changes nothing in your workspace. Capacity confirmation only comes up when you start a discovery run or qualify a new post's audience. Tracking a new profile also runs an initial discovery, and the qualified leads it delivers use your workspace capacity (Tracking docs).

What you can read: performance, post analytics, leads per post, meetings

The traxy MCP has a Content, analytics and meetings tool group, plus lead and Tracking tools that help with reporting. Here is what each one reads, in the words of the MCP tool catalog:

A few notes on reading this table:

• Performance is the chat version of Insights in the app, which shows lead volume, recent trends, audience breakdowns and the top profiles or posts bringing in leads (Insights docs).
• Post analytics covers posts from the profiles you track.
• Leads per post lists the people behind the count. Each lead comes with their ICP match and the signal behind them, so you can see why a post counted.
• Meetings are calendar meetings that traxy matched to a lead or the lead's company. The outcome on each one (Active opportunity, Already a client or Closed won) is recorded in the app under Outbound > Meetings (Meetings docs).

For the full catalog by job, including the outreach and inbox tools, read what the traxy MCP server can do.

Prompts for a weekly summary

A weekly summary is the most common use: one message on Monday that tells you what last week's posts and outreach brought in. Copy these prompts and replace the parts in braces. What comes back depends on your workspace.

A one-message weekly summary

In {workspace}, summarize the last 7 days. Use the analytics dashboard for overall performance, list the tracked-profile posts from that period, and for the three posts with the most qualified leads, list those leads with role, company and ICP match. Then give meeting counts and the attributed meetings for the same period. Report qualified leads, campaign results and meetings as separate sections, and name the tool behind each number.

A short version for Slack or email

Turn that summary into five lines I can paste into our team channel: qualified leads this week, the best post and why, the best tracked profile, meetings attributed, and one thing to change next week.

Week over week

Compare the last 7 days with the 7 days before in {workspace}: qualified leads, posts that brought in at least one qualified lead, and attributed meetings. Use the same scope for both weeks and tell me if any number can't be compared.

Asking for separate sections matters. The Meetings docs put it plainly: a delivered message is not a reply, and a reply is not a booked meeting. A good summary keeps each stage on its own line instead of blending them into one score.

Save the prompt you like and run it every week with the same period and scope, so the weeks compare. For a routine that also finds new leads every morning, see how to run a daily LinkedIn lead search with an AI agent.

Which posts and tracked profiles brought qualified leads?

Ask the AI to rank posts by the qualified leads they brought in, then open the top ones to see the people. The useful answer is a short list of posts with names behind them, not a total.

In {workspace}, list the tracked-profile posts from {period}. For each, give the number of qualified leads from traxy_list_post_qualified_leads. Rank them, and for the top five show the leads with role, company, ICP match and whether they liked, commented or reposted.

Show profile totals for each tracked profile in {period}. Which competitor, creator or teammate brought the most qualified leads, and which brought none?

For {post link}, read the post summary and list its qualified leads. What do the leads have in common (roles, company size, industry)?

What you do with the answer:

• Plan content. Topics whose posts bring in people who fit your ICP are worth repeating. Posts with high engagement and no qualified leads reached the wrong audience, however good the numbers look on LinkedIn.
• Review Tracking. A tracked profile that brings no qualified leads over several weeks is a candidate to pause. One that keeps bringing people who fit is a source for a campaign of its own. You can pause, remove or restore tracked profiles from the same chat.
• Check the leads themselves. Before you act on a count, read a few leads and the comments behind them. If matches are weak, the Insights docs say to inspect individual leads and refine targeting before changing anything else.

In the app, the same view lives under Targeting > People (open a profile) or Targeting > Posts: each tracked profile shows the leads it surfaced, their ICP match and the posts they came from. For how to choose who to track in the first place, read who to track on LinkedIn to find leads.

Building a client report from chat (agencies)

An agency can build each client's report from one chat: grant the connection access to every client workspace, name the workspace in each request, and ask for the same report shape every time.

Run each client in its own traxy workspace, with its own ICP, Tracking and sender accounts. The MCP works on the default workspace you authorized or on the one you name, so the risk is mixing clients up. Naming the workspace in every prompt avoids that.

In {client workspace}, build this month's LinkedIn report for {client}. Sections: (1) posts from {client}'s tracked profile, ranked by qualified leads; (2) the qualified leads from the top posts, with role, company, ICP match and the post they engaged with; (3) campaign results from outreach to those leads; (4) attributed meetings, with the matched lead or company. Keep qualified leads, replies and meetings in separate sections. Add a one-paragraph summary at the top in plain English for a non-technical client.

Then shape it for the client:

• Brand and format. Ask the AI to lay the report out with the client's name and your agency's section order, as a table or a document you edit before sending.
• Explain fit. For each lead, ask for one line on why they fit the client's ICP, taken from the lead's data, so the client sees the reasoning behind each name.
• Check outcomes in the app. Meeting outcomes are recorded under Outbound > Meetings in each client's workspace. Record them there before the report, so the meeting list the AI reads is current.
• Repeat per client. Run the same prompt with the next workspace name. Same shape, same period, so clients and months compare.

If you prefer the report in the app, with exports and a weekly lead list to the client's Slack, read how to show clients which LinkedIn posts brought in leads. For proving meetings at renewal time, read how to track meetings booked from LinkedIn outreach. Agencies can run traxy for their clients as a managed service; see traxy for agencies.

What happens after the report: outreach from the same chat

The report tells you who your posts and tracked profiles brought in. The same MCP connection can then run outreach to them: every traxy customer can draft, test and launch campaigns, manage sequences and senders, and reply in the inbox from Claude or ChatGPT, with the AI asking for your yes before anything sends or launches.

1. Pick the audience from the report. For example, the qualified leads from your three best posts this month, or everyone a competitor's posts brought in.
2. Draft a campaign. The campaign researches each lead for the facts you choose, groups leads by signal and prepares a personalized message variant for each group, over LinkedIn and email.
3. Test on sample leads. Review the research and drafts with nothing sent.
4. Review the launch and say yes. See who joins and who is skipped, then approve. It sends from your own connected LinkedIn and email accounts. You can let the approved campaign send, or approve each message in a review queue.
5. Let replies stop it. Replies, unsubscribes and hard bounces stop the sequence, and replies land in one inbox you can also answer from chat.

Next week's summary then shows how that campaign did. Because each campaign can start from a different signal, you can run several at once and compare them in the same report. For the outreach side in more detail, see LinkedIn outbound with traxy, and for the finding side from chat, read how to find and research LinkedIn leads from Claude.

What the numbers mean, and what they don't

The numbers are counts of events in your workspace. They are useful for comparing posts, profiles and weeks, and easy to misread if you add them up or treat them as revenue.

• Different events are different numbers. The Insights docs say a profile scan, a post engagement and a delivered qualified lead are different events. Engagement on a post is not a lead count.
• Post counts don't add up to a lead total. Several posts can include the same person, so adding post-level counts may not equal your unique lead total. If you want "how many people", ask for unique leads, not the sum of posts.
• A qualified lead is a fit check. It means the person showed social intent and has a 50%+ match to your ICP. It does not mean they want to buy, and the fit score is not a probability of purchase.
• Delivered, replied and booked are separate stages. A delivered message is not a reply, and a reply is not a booked meeting. Report each one on its own line.
• Meetings depend on matching. Attribution uses the attendee and lead information available. Connecting a calendar doesn't mean every event will match, and a meeting can match the lead or only the lead's company. If an expected meeting is missing, check calendar selection and sync state.
• An attributed meeting is not a closed deal. The outcome you record (Active opportunity, Already a client, Closed won) says what the meeting became. Revenue belongs in your CRM.
• Keep the scope consistent. Compare numbers only with the same workspace, date range and filters. The docs recommend keeping the same scope when you move between Insights, Tracking and Leads, and the same holds in chat.

The safest habit is to ask the AI to show its sources: which tool, which workspace, which period. If a number surprises you, ask for the leads or meetings behind it rather than the total.

What are the limits of reporting from chat?

• Your permissions decide what you see. Scopes, workspace role and product access filter what tools/list returns. If a reporting tool is missing, reauthorize with the analytics scope or check your role.
• Post analytics covers tracked profiles. Posts from profiles you don't track aren't in it. Track a profile first, then report on it.
• Meetings need a connected Google Calendar. Outcomes are set in the app under Outbound > Meetings, not from chat.
• The AI summarizes, you check. Claude or ChatGPT can mislabel a number or add counts that shouldn't be added. Ask for the tool behind each number and check anything that goes to a client.
• Contact details follow permissions. Contact permissions control attendee detail access for meetings, so a report may show a meeting without the attendee's contact details.
• It reports what traxy records. Clicks, deals and revenue live in your website analytics and CRM, so take those numbers from there.

FAQ

Does ChatGPT work the same way as Claude?

Yes. The traxy MCP connects to Claude, ChatGPT, Codex, Cursor and Grok with the same URL and a one-time sign-in, and the reporting tools are the same in each.

Do I need an API key?

Not for the MCP. You sign in to traxy once when you connect. If you'd rather pull the numbers into your own code, the traxy REST API reads leads, signals, tracked profiles, analytics and meetings with an API key from Integrations.

Can Claude see my LinkedIn post impressions?

Claude sees what traxy reads: aggregate performance, post analytics for tracked profiles, metrics for one post and the qualified leads each post brought in. Ask it to list the metrics a tool returned for your workspace rather than assuming a field exists. If you need a metric traxy doesn't return, LinkedIn's own post analytics is the source.

Does asking for a report use lead capacity?

Reading reports doesn't start a discovery run. The reporting tools read data you already have. Starting a discovery run or qualifying a new post's audience can add qualified leads, and the AI asks you to confirm capacity use before it does either. Tracking a new profile also runs an initial discovery, and the qualified leads it delivers use your workspace capacity.

Can I see which leads from a post booked a meeting?

Yes. List the qualified leads for the post, then ask the AI to check them against the attributed meetings for the same period. Each meeting opens to its matched lead or company, and each lead keeps the post or comment behind it.

Can the AI send the report to my client automatically?

The MCP reads the numbers; how you send the report is up to you and your AI client. In the app, traxy can route qualified leads to a client's Slack channel. Review anything client-facing before it goes out.

Can I report on several client workspaces in one chat?

Yes, if each workspace is granted to the connection. Name the workspace in every request so the numbers come from the right client, and build one report per client rather than mixing them.

Can I launch outreach to the leads in the report from the same chat?

Yes. Draft a campaign for those leads, test it on sample leads with nothing sent, review who joins, then say yes to launch. LinkedIn and email steps send from your own connected sender accounts, and replies, unsubscribes and hard bounces stop the sequence.

Related reading

• The LinkedIn GTM MCP and LinkedIn outbound with traxy
• How to show clients which LinkedIn posts brought in leads
• How to track meetings booked from LinkedIn outreach
• How to find and research LinkedIn leads from Claude
• What the traxy MCP server can do, by job