X Profile Reactions
/x-profile-reactionsRank any X account's posts by the likes, retweets, replies and views each one drew, with date range and sort filters.
Overview
x-profile-reactions calls the SyncGTM MCP x_profile_reactions tool to return an X profile's posts with the reaction totals each one drew: likes, retweets, replies and views. It adds filters the plain post tool does not have, including a date window, sort order, extra handles in the same call, and search terms. It gives you per-post totals, not the accounts behind them, for 0.2 credits per call.
What it does
- 1
Takes one full X profile URL. profile_url is required and a bare handle will not resolve.
- 2
Returns each post's text with its likes, retweets, replies and views as per-post totals. The response is text plus those four numbers — there is no post URL field in it.
- 3
Filters the plain post tool lacks: start_date and end_date in YYYY-MM-DD, sort set to Top, Latest or Latest + Top, and max_posts from 1 to 100 (default 10).
- 4
Folds several accounts into one call with twitter_handles, written without the @, so a four-competitor set costs the same as one profile.
- 5
Narrows to a theme with search_terms, but only when include_search_terms is set to true. The default is false and the terms are otherwise ignored.
- 6
Does not return the accounts that reacted, follower counts, emails or firmographics. You get the number 214, not the 214 handles behind it. Costs 0.2 credits per call whatever max_posts is set to.
How to use it
- 1
Download the skill file — it saves as x-profile-reactions.zip.
- 2
In Claude, open Settings → Capabilities → Skills and upload the .zip (or unzip the x-profile-reactions/ folder into .claude/skills/ for Claude Code).
- 3
Connect the SyncGTM MCP server — browser sign-in, no API key — then run check_credits, which is free, to see your balance.
- 4
Trigger it by typing /x-profile-reactions and pasting a full profile URL such as https://x.com/stripe.
- 5
Tune one setting: max_posts. It costs the same at 100 as at the default 10, so set 50 or 100 unless you want a spot check.
Use cases
Competitor message testing
See which of a competitor's posts actually earned replies and retweets. Borrow the angle their audience rewarded instead of the one they repeat most.
Founder outreach openers
Pull a founder-led prospect's best-performing post and open with the argument inside it. Useful when the buyer is quiet on LinkedIn but active on X.
Launch-week benchmarking
Window start_date and end_date around your launch and compare your post totals against three competitors folded into the same call.
Content brief evidence
Rank a category leader's posts by views and hand your content team the themes that outperformed, with numbers attached rather than opinions.
GTM workflow examples
Monday competitor teardown into a warm reply list
- 1Call `x_profile_reactions` on your closest competitor's X profile with sort "Top", max_posts 100 and start_date set to the first day of last quarter — 0.2 credits for up to 100 posts with their likes, retweets, replies and views.
- 2Fold the other three competitors into that same `x_profile_reactions` call with twitter_handles (no @), then compute each account's median likes so posts from a 90k-follower account and a 9k-follower account are compared fairly.
- 3Take the two highest-reply posts per account and copy their URLs off X — `x_profile_reactions` returns post text and counts, not links — then run `x_post_commenters` on each URL at 0.4 credits per call with max_results 100 to pull the handles arguing in the replies.
- 4Match the commenters whose bio names an employer to LinkedIn with `find_people` at 0.3 credits per result, limit 25 for 7.5 credits, then run `find_work_email` at 1 credit on the ICP matches only.
Outcome: A one-page teardown naming the three message angles that outperform in your category, plus a 25-row lead list of people who publicly engaged with those angles.
Pre-call brief for a founder-led account
- 1Run `x_profile_reactions` on the founder's profile with sort "Latest", max_posts 50 and a start_date 90 days back — 0.2 credits for their recent posts and the totals each drew.
- 2Re-run `x_profile_reactions` with search_terms ["hiring", "pipeline"] and include_search_terms true to confirm whether the problem you sell into shows up in their own words — another 0.2 credits.
- 3Run `linkedin_profile_posts` at 0.3 credits on that person's LinkedIn URL, which you supply because the X tool does not carry one, and note where the two channels say different things.
- 4Run `linkedin_profile_enrich` at 1 credit on the same LinkedIn URL for their work experience, education and skills — the history an X profile does not carry.
Outcome: A half-page call brief quoting the two posts that beat the founder's own median, their recurring topic, and an opener built on the argument instead of the job title.
Launch-week share of voice scorecard
- 1Run `x_profile_reactions` on your own X profile with start_date and end_date bracketing launch week, sort "Latest + Top" and max_posts 100 — up to 100 posts inside that window with likes, retweets, replies and views for 0.2 credits.
- 2Add the three competitors who shipped in the same window to that `x_profile_reactions` call via twitter_handles, so the whole comparison set costs nothing extra.
- 3Re-run `x_profile_reactions` over the same window with search_terms ["pricing"] and include_search_terms true to separate the pricing conversation from general launch noise.
- 4Copy the URLs of the two posts that drew the most replies from X and send each to `x_post_commenters` at 0.4 credits per call, max_results 100, to check whether the replies came from buyers or from peers.
Outcome: A launch scorecard comparing your posts against three competitors line by line, with a call on which post format to repeat next launch and whether it reached buyers.
Prompts
Paste any of these into Claude once the skill is installed and the SyncGTM MCP is connected.
Get reactions for the X profile https://x.com/stripe with sort Top and max_posts 50. Rank the posts by likes and flag the ones that beat the account's median.Pull x_profile_reactions for https://x.com/ramp between 2026-06-01 and 2026-07-31 with sort Latest and max_posts 100, then group the posts into 3 themes with combined likes per theme.Run x_profile_reactions on https://x.com/ramp and also include twitter_handles brex, mercury and stripe. Set max_posts 50 and sort Top, and give me one table ranked by likes with the account in the first column.Use x_profile_reactions on https://x.com/vercel with search_terms ["pricing"], include_search_terms true, sort Top and max_posts 50, so I only see posts about pricing.Run x_profile_reactions on https://x.com/stripe with max_posts 50 and rank the posts by replies. It returns no post URLs, so list the top 2 and ask me to paste their links, then run x_post_commenters on each with max_results 100 and shortlist commenters whose bio names a B2B SaaS company.Best practices
0.2 credits per call no matter what max_posts is, so set 50 or 100 on the first call. Leaving the default at 10, seeing too little and calling again is how you pay twice for one answer.
Batch accounts with twitter_handles instead of one call per handle. Four competitors in one call is 0.2 credits; four separate calls is 0.8 for the same table.
search_terms only applies when include_search_terms is true, and the default is false. Set both or the filter is silently ignored and you get an unfiltered feed.
Pass the full profile_url. A bare handle is not enough here, unlike the extra handles you add through twitter_handles.
Use sort Latest for recency and Top for signal — Top is the default. Add start_date and end_date whenever the question is about a launch or a campaign, or you get whatever the sort order happens to surface.
Ask for reaction counts, never for who reacted, and do not expect a post URL back — the response is post text plus the four counts. To get the accounts behind a number, copy that post's URL from X and run x_post_commenters at 0.4 credits per call. There are no emails anywhere in this response.