creator-vetting

Creator vetting

A go or no-go on a creator, from public numbers: how many people really see their posts, how those people react, how steady they are, how much they already sell, and anything a brand should worry about.

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.

Gather

For each platform the creator is on (a handle or a profile link is enough):

TikTok

• tiktok_profile(handle=<handle>): stats.followerCount, stats.heartCount, stats.videoCount, user.verified, user.bioLink, and user.privateAccount (private means stop: nothing to vet).
• tiktok_profile_videos(handle=<handle>) twice (the second with the first page's max_cursor) for about 30 posts: stats.playCount, stats.diggCount, stats.commentCount, stats.shareCount, createTime, isAd and desc each.

Instagram

• instagram_profile(username=<username>): data.follower_count, data.media_count, data.is_verified, data.category, data.external_url.
• instagram_user_posts(username=<username>, first=30): like_count, comment_count, play_count (videos), taken_at, is_paid_partnership and caption per post. A hidden like count comes back null; say so rather than counting it as zero.

YouTube

• youtube_channel(handle=<handle>): subscriberCount (rounded by YouTube), viewCount, videoCount, joinedDate, country, links.
• youtube_channel_videos(handle=<handle>, sort=latest): about 30 videos with viewCountInt and publishedTimeText. Shorts are separate: youtube_channel_shorts(handle=<handle>).

Their other accounts: the bio links above, or linkinbio_linktree(url=<linktree url>) when the bio points to a Linktree (its socialLinks and links show every account and current brand deals).

Their audience: comments on two recent typical posts (not the viral one): tiktok_comments(url=<video url>), instagram_post_comments(url=<post url>) or youtube_comments(url=<video url>).

Brand safety, if asked or if anything looks off: the transcripts of two or three recent videos, tiktok_transcript(url=<video url>) or youtube_transcript(url=<video url>).

Compute

Work from the numbers, not impressions:

• Median views of the last 30 posts (Instagram photo posts: median likes plus comments). Use the median: one viral post wrecks an average.
• Reach rate: median views divided by followers. It's the best single check on bought or dead followers.
• Engagement rate: (likes + comments + shares) / views per post on TikTok and YouTube; (likes + comments) / followers on Instagram. Take the median.
• One-hit check: mean views over median views. Above about 3, the average is carried by a few hits.
• Rhythm: posts per week over the window, and the longest gap.
• Sponsorship load: share of posts with isAd, is_paid_partnership, or #ad, #sponsored or "paid partnership" in the caption, and which brands.
• Audience quality: the share of comments that are generic ("🔥", "nice", emoji only), in an unexpected language, or repeated; versus real questions and replies.

Red flags

• Reach rate far below the creator's peers of the same size, or follower jumps with flat views.
• Mean several times the median: the "average views" in their media kit won't repeat.
• Comments that don't match the content, or a high share of generic ones.
• Many sponsored posts in a row, or a direct competitor among recent partners.
• Long gaps, or a falling trend over the last 30 posts.
• Anything in captions or transcripts the brand wouldn't want next to its name.

Budget

One platform is about 3 to 5 calls (profile, two pages of posts, a page or two of comments); a creator on two platforms is about 8 to 10, plus transcripts if checked. For a shortlist, look up every profile first (one call each), then vet only the top few in depth; say the estimate before a long list.

Report

# Creator vetting: @<handle>, <date>

| Platform | Followers | Median views | Reach rate | Engagement | Posts/week | Sponsored |
| --- | --- | --- | --- | --- | --- | --- |

**Verdict:** Go / Go with conditions / Pass, in one sentence.

**Strengths.** 2 to 3 bullets with numbers.
**Risks.** Each red flag found, with the evidence and a link.
**Audience.** What commenters ask and say, with two quotes.
**Their best format.** What the top posts have in common.
**If you go ahead.** Deliverables and metrics to ask for, from what the data shows.

Pitfalls

• Compare a creator with their own baseline and with peers of their size, never with raw follower counts.
• Views on recent posts are still climbing: leave out posts younger than about three days, or say they are.
• YouTube subscriber counts are rounded; Instagram may hide likes.
• Public data shows reach and reaction, not audience demographics or conversions. Say what you couldn't check.

Calls and credits

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