Lurk a brand

One name or website in, one page out: is the brand advertising right now, where, saying what, and how big is its audience. Four calls in two rounds. For a full teardown of the ads afterwards, use the ad-research skill.

Calling LurkAPI

With MCP (the LurkAPI plugin or connector), call the tools by name, like tiktok_profile; your client may prefix them (mcp__plugin_lurkapi_lurkapi__tiktok_profile). With only find_endpoints and call_endpoint, run a tool as call_endpoint with endpoint set to its name and params to its arguments. Results come trimmed: no nulls, one URL per image or video. A long result comes in parts: its _note gives the offset for the next part, which is free.

Without MCP (Codex, Cursor, scripts), every tool is GET https://api.lurkapi.com plus its path, with the same params as a query string (URL-encoded) and the key from LURKAPI_API_KEY in the x-api-key header:

curl -s "https://api.lurkapi.com/v1/tiktok/profile?handle=nike" -H "x-api-key: $LURKAPI_API_KEY"

If LURKAPI_API_KEY isn't set but LURKAPI_KEY is (the name older docs used), use that. Raw answers are big: pipe them through jq to keep the fields you need. No key yet? Ask the user to create one at https://lurkapi.com/dashboard (100 free credits on sign-up, then a balance below 25 tops back up to 25 every day) and to set LURKAPI_API_KEY in their shell. Never ask for the key in the chat, and never print it.

Credits. Each call costs what the table at the end says, not_found included; failures on our side (5xx) are free. Every answer carries credits_remaining. Before a run of more than 20 calls, check the balance with account_credit_balance (free) and tell the user what the run will cost. If a call answers missing_api_key, insufficient_credits or signup_required, stop and pass on what it says.

Run

Work out the brand's name (Nike) and its website (nike.com) from what the user gave. If you're unsure of the website, ask, or try <name>.com and say so.

Round 1, both at once:

1. facebook_search_companies(query=<name>) finds the brand's Facebook page. Take the first of searchResults unless a later one is clearly the brand: the exact name, verification of BLUE_VERIFIED, the most likes. Big brands have regional and product pages; name the runner-up if it was close. Keep its page_id, likes (Facebook followers), ig_username and ig_followers.
2. google_company_ads(domain=<website>, region=US) lists the ads promoting the site on Google Search, YouTube, Maps, Play and Shopping. Read number_of_ads_estimate, and from ads the advertiserName, format and firstShown of the first few. Anyone advertising the site shows up here, resellers and agencies included; the ads are the brand's own only when advertiserName contains its name. Use the brand's home country as region when it isn't American.

Round 2, both at once, as soon as the page is known:

3. facebook_company_ads(pageId=<page_id>, status=active, trim=true) returns the live ads on Facebook and Instagram, most seen first, up to 30. searchResultsCount is Meta's estimate of how many are live.
4. tiktok_profile(handle=<ig_username>): brands use the same handle on both far more often than not. Read stats.followerCount, stats.heartCount and stats.videoCount. Check it's really theirs: an account with no bio (user.signature), a handful of followers and no videos is a squatter, so say the brand doesn't own the handle. A not_found means there's no TikTok under that handle; say so rather than guessing another.

No page found? Skip both round 2 calls, which need it, and say the brand has no Facebook page under that name. No ig_username? Skip the TikTok call.

That's the lurk: four calls. Stop there unless the user asks for more; each extra lookup (instagram_profile, youtube_channel, linkedin_company) is one more call.

Read the ads

From the live Meta ads, find what repeats, not what's listed first:

• The offer and the angle. The primary text is snapshot.body.text (or snapshot.cards[].body for carousels); the headline is snapshot.title. Quote real lines.
• The ads that work. An ad still running weeks after its start_date (Unix seconds) is the best public sign that it pays. A high collation_count means many versions of one ad: they're scaling or testing it.
• Where it sends people. snapshot.link_url and snapshot.cta_text.
• Where it runs. publisher_platform: Facebook, Instagram, Messenger, Threads.

Skip ads whose text holds template tags like {{product.name}}: those are catalog ads and say nothing about the brand's message.

Report

## <Brand>, lurked <date>

| | |
| --- | --- |
| Meta ads live | about <n> (Facebook page <name>, <likes> followers) |
| Google ads | about <n>, as <advertiser name>, mostly <format> |
| Instagram | @<handle>, <followers> followers |
| TikTok | @<handle>, <followers> followers, <likes> likes |

**What they're pushing.** 3 to 5 bullets: the offers, products and hooks that repeat in the live ads, each with a quoted line.

**The ad to study.** The longest-running live ad: its copy, how long it has run, where it sends people, and its link (https://www.facebook.com/ads/library/?id=<ad_archive_id>).

**Read.** Two sentences: how hard they're advertising, and where they're strongest.

Pitfalls

• Counts are estimates (Meta's and Google's). Say "about".
• Live doesn't mean winning: only run time hints at that. Spend and results aren't public, except for political ads.
• Google's region=US counts only ads shown in the US; ALL counts everywhere.
• The Facebook page search can match a fan page or a namesake. If the followers look wrong for the brand, say so before reading its ads.

Calls and credits

Each tool's link opens its docs as markdown: every param and its values, and a real response.