- You may be asked to provide your CreatorDB API key the first time you run a query with a skill. This is required for authentication when Claude Code calls the CreatorDB API.
- Check the CreatorDB API documentation for details on the available endpoints, request parameters, and response formats to better understand how to structure your queries and interpret the results.
- Check the API usage and pricing details to understand the cost implications of your queries and manage your credits effectively. For details, see API Credit Usage.
Responses below are illustrative reconstructions, not verbatim terminal output — exact wording, table layout, and credit numbers will vary across runs.
Example 1: Check your account balance
To check your CreatorDB account balance, you can use the account skill. For example, you can prompt Claude Code with the following query:
Prompt example
Response example
Claude Code will then use the account skill to call the CreatorDB API and retrieve your account information. Here’s your CreatorDB account snapshot.Balance
- Credits remaining: 989,923
- This
/usage call cost 1 credit.
Activity (last 7 days)Days 5/1–5/3 had no activity (omitted because the API doesn’t return zero-usage days).Where credits are goingThree contact-and-audience endpoints account for 74% of spend:Platform mix
- YouTube: 7,946 cr (81.7%)
- Instagram: 1,376 cr (14.1%)
- Sponsor endpoints: 404 cr (4.2%)
- TikTok: 0
- Account (
/usage): 1 cr (<0.1%)
Example 2: Search for a specific creator
To search for a specific creator, you can use the creator-search skill. For example, to find information about the YouTube creator “PewDiePie”, you can prompt Claude Code with the following query:
Prompt example
Response example
Claude Code will then use the creator-search skill to call the CreatorDB API.
- Category mix: Gaming 51.5% · Entertainment 48.3% · People & Blogs 0.2%
- Topics: social-media comedy, variety gaming, vlogs, memes
- Niches: meme/gaming, Clash of Clans, Reddit reactions, art/drawing, Japan, KSI-related, fitness, “pewdiepie” niche (eponymous)
- Subscriber growth (7d/30d/90d): all flat at 0 — channel has plateaued at the 110M ceiling
- Recent hashtags hint at sponsors: #ad, #partner, #shopad, #clashofclans, #clashfornature, #gfuel, #honor20pro — Clash of Clans, G FUEL, Honor (Huawei sub-brand) all visible
- Instagram: @pewdiepie
- TikTok: @pewdiepie
- Twitch: pewdiepie
Estimated rate card (CreatorDB model — benchmark, not a quote)
Top 15 Related Creators (of 323 returned)
- Top of the related list = closest co-watch / similar-audience peers. The first three (Markiplier, penguinz0, Veritasium) reflect PewDiePie’s actual audience overlap — gaming-comedy + science creators his viewers also subscribe to.
- MrBeast and MrBeast Gaming sitting at #9–10 makes sense given collab history and audience overlap.
- Late-night TV brands (Daily Show, Last Week Tonight, Jimmy Kimmel) appearing high in the related list is interesting — likely reflects English-speaking general-entertainment co-viewing, not direct topical similarity.
- Technoblade (#14) — note this is a memorial channel; the creator passed in 2022 but the channel remains active.
- The full 323-creator list is available if you want the long tail (would cost ~616 cr to enrich them all to names; happy to enrich any specific batch you point at).
Credits used this turn: 32 (1 PewDiePie profile + 15 related profiles + 1 prior balance probe was already counted). Remaining: ~991,777.Example 3: Search for creators that meet specific conditions
To search for creators based on specific criteria, you can use the creator-search skill. For example, to find Instagram creators in the US “gaming” category with more than 1 million followers and high engagement rates, you can prompt Claude Code with the following query:
Prompt example
Response example
Claude Code will then use the creator-search skill to call the CreatorDB API. Top 10 — US Gaming Instagram, 1M+ followers, ranked by Reels engagement rate
Here’s the ranked top 10. Worth flagging some quality caveats below.
- Confirmed core-gaming creators (8 of 10): shroud, Sketch, Nadeshot, ronnie2k (NBA 2K), and the four esports orgs (NRG, LoL Esports, G2, plus the “Gamer”-categorized annashumatee). These are unambiguously gaming.
- Verify before outreach (2 of 10):
- #1 @nihaachuu (6.98% ER) — IG category is “Video creator”, not “Gamer”. She showed up in the gaming-niches search but her primary content is modeling/lifestyle, not gameplay. ER is genuinely high but she may not be a fit if you specifically want gaming-content creators.
- #3 @thepointerbrothers — also “Digital creator”. They make NBA / sports-comedy content; appeared here because they used #gameday / #gaming hashtags. Almost certainly not a gaming creator.
- Industry benchmarks: <0.5% is typical for mega-IG accounts, 1–3% is healthy mid-tier, >5% is exceptional. The whole list lands in the healthy-to-exceptional range, which is plausible for gaming-niche IG (smaller, more loyal audiences).
Instagram’s gaming discovery in CDB is hashtag-driven and noisy:
- The id_gaming_All niche pulled 23K creators including HOT 97, Demi Lovato, Sanrio, Matthew Lillard.
- NLS for the same query surfaced Zendaya, Cardi B, Rihanna alongside real gaming creators.
- The IG category field (e.g., “ESPORTS_TEAM”, “Gamer”) is more reliable than niches but only ~5–10% of creators have a gaming-specific value.
So this list is the result of: niche-search candidates → manual filter to gaming-clear handles → enrich with /instagram/performance for actual ER → rank. If you need a strictly-vetted list, drop @nihaachuu and @thepointerbrothers and pull the next two candidates (@dakotaz at 1.49% ER and a couple from a wider search).
Search probes + 12 × profile + 12 × performance ≈ ~60 cr total. Remaining: ~988,200.
Example 4: Build an outreach list end-to-end
To create an outreach list of potential creator partners for a campaign, you can use a combination of the creator-search and creator-enrichment skills. For example, to find 50 beauty TikTokers with more than 500,000 subscribers in the US with high engagement rates and retrieve their contact information, you can prompt Claude Code with the following query:
Prompt example
Response example
Claude Code will then use the creator-search and creator-enrichment skills to call the CreatorDB API. Top 20 — US Beauty TikTok, 500K+ followers, ranked by engagement rate
Rows 21–50 are saved in /{YOUR WORKING DIRECTORY}/{FILENAME}.csv with full bios, niches, and verification flags. Notable agency contacts in the long tail: kensingtongrey, dulcedo, undercurrent.net, viralnationtalent, fixated, lucky8collective, thedigitalbrandarchitects.
- The ER values are TikTok-formula (likes + comments + shares) / followers, which runs noticeably higher than YouTube/Instagram ER. #1 @cannolifamily at 97.05% is almost certainly a recent-window artifact (one viral video pulling the rolling average up); double-check before reporting it as a sustained metric. #2 @michaelahallmark1 at 34% is also extreme — verify content fit (her bio mentions “MakeJesusKnown” hashtags, may be more lifestyle/faith than beauty).
- Niche-tag noise: The TikTok beauty niches are hashtag-driven, and several handles in the list don’t clearly skew beauty:
- @still.bad.decisions (#20) — bio says “TheCarGirl BackUp Cars, corgs, and complaining”
- @dunknatachai (#34, 3.2M) — Thai actor (Disney/GMMTV credits)
- @sway_bhatia1 (#31) — actress
- @christwodafur (#38, 3.1M) — Star Wars cosplay
- Some have very generic bios suggesting general lifestyle content rather than beauty
The strongly-beauty-confirmed handles are the ones with agency emails (kensingtongrey, dulcedo, etc.) and bio mentions of makeup/skincare. If you want a strictly beauty-only roster, filter the CSV to rows with beauty-keyword bios or pull /tiktok/sponsorship to confirm beauty-brand work history.
Bio-first contact approach saved ~405 cr vs. blindly calling /contact on all 50. Remaining: ~987,650. Examples 5a-5c each show a different cut of the brand-sponsorship skill: known brand → roster (5), vertical → brands (6), brand → spend estimate (7).
brand-sponsorship skill to analyze a brand’s sponsored creator roster, identify which brands are active in a specific vertical, and estimate a brand’s monthly spend on creator partnerships.
To verify a brand’s sponsored creator roster, you can use the brand-sponsorship skill. For example, to find out which creators are sponsored by Adidas, you can prompt Claude Code with the following query:
Prompt example
Response example
Claude Code will then use the brand-sponsorship skill to call the CreatorDB API.
YouTube — geographic spread
Top 10 countries by creator count: USA 69 · ESP 25 · JPN 23 · GBR 17 · FRA 17 · THA 17 · POL 16 · KOR 16 · BRA 15 · DEU 15True global program — not US-centric. Spain, Japan, and Thailand are notably high (football market + Asian apparel push).
Most-active YouTube partners (by # placements with Adidas)
Pattern: Adidas leans heavily on mid-tail soccer/football review channels for repeat partnerships (boots, kits, gear reviews) rather than mega-creators. Volume of placements doesn’t track follower count — Fotbollsmorgon has 25 placements at just 40K subs, while 12.9M-sub J Jai Pan has only 1. Largest YouTube creators (by reach)
12.9M (J Jai Pan, THA) · 8.57M (loveliveserve, USA) · 6.91M (Complex, USA) · 5.40M (Unisport, 7 placements — also a heavy partner) · 5.40M (Blowek, POL) · 2.89M (DAZN Football) · 2.32M (adidas’ own channel — interesting that it shows up as a “sponsored creator” of itself, treat as a self-reference noise row).
Only 27 IG creators indexed. Top by reach: N’Golo Kanté (16.5M, 2021), Trinity Rodman (750K, 2026-01), Melanie Leupolz (290K, 2024). Most have just 1 placement and many entries are very small accounts (<50K). The IG sponsored-content index for Adidas appears thin compared to YouTube.
- adidas channel itself appears in the YouTube list (USA, 2.32M subs, 1 placement). That’s the brand’s own channel — likely a self-tag from a user-generated content collab. Filter it out for clean creator-side analysis.
- CreatorDB sponsorship detection is heuristic — undisclosed paid partnerships and creators using non-standard tags are missed. Real Adidas roster is likely larger.
- The 27 Instagram results are a suspiciously low count for a brand of Adidas’s scale — IG sponsorship indexing in CDB is generally less complete than YouTube’s. If IG is critical, treat this list as partial.
Remaining: ~990,189.CSV columns: platform, creatorId, displayName, country, followers, sponsoredCount, lastSponsoredDate, avgRecentSponsoredEngagementRate, topics, niches. Open from your working directory with Numbers or Sheets. brand-sponsorship skill. For example, to find the top 10 brands sponsoring US gaming YouTubers, you can prompt Claude Code with the following query:
Prompt example
Response example
Claude Code will then use the brand-sponsorship skill to call the CreatorDB API. Top 10 Brands Sponsoring US Gaming YouTubers
- Game publishers and individual game titles dominate — 7 of the top 10 are gaming companies promoting their own games. Epic Games leads on roster reach (Fortnite-related promos especially), Ubisoft leads on placement volume (15 placements across 4 creators = repeat partnerships).
- Mobile gacha / F2P MMOs are aggressive sponsors — Marvel Rivals, AFK Journey, Crystal of Atlan, War Thunder all rely heavily on creator marketing for user acquisition. NetEase appears under multiple titles (Marvel Rivals + Crystal of Atlan).
- Non-gaming partners: Chime (banking) and NordVPN (VPN) — the two evergreen “every creator” sponsors. Chime’s 8 placements across just 2 creators shows it’s a high-frequency repeat partner where it does land.
- Hardware tier: Intel (3) and Alienware (2, just outside top 10) — the consistent PC-hardware pair.
- Just outside top 10: Alienware, 2K, Lurkit (a creator-marketing platform itself), GeoGuessr, my.games — all at 2 creators each.
- Sample of 50 creators, sorted by subscribers descending — top-of-funnel skew. Smaller channels in the long tail may show different sponsor patterns.
- CreatorDB sponsorship detection is heuristic (disclosure tags, hashtag conventions). Undisclosed paid placements get missed; the actual ranking could shift slightly.
- /youtube/sponsorship returns up to ~20 most recent sponsored uploads per creator, not lifetime — so very old partnerships won’t appear.
- Tie-breaking near the cut: brands ranked #8–10 all have only 2–3 creators in the sample. A larger sample (e.g., 200 creators) would give more robust ranks below the top 5.
- Expand the sample to 200 creators (~1,000 cr more) for tighter ranking below the top 5
- Filter to a specific gaming sub-niche (e.g., FPS, Roblox, Minecraft) using the niches filter
- Cross-reference with /sponsor/summary for the top brands to see their estimated 30-day spend on YouTube overall
- Get the audience demographics of each top brand’s sponsored pool with /sponsor/audience
Credits used this turn: ~251 (1 search + 50 sponsorship). Remaining: ~991,558.
brand-sponsorship skill. For example, to find Nvidia’s latest sponsoring activities, you can prompt Claude Code with the following query:
Prompt example
Response example
Claude Code will then use the brand-sponsorship skill to call the CreatorDB API. Nvidia — estimated influencer spend, last 30 days
~$24,448 USD (YouTube only; CreatorDB model estimate)
- The 90-day total is
$249K, so the last 30 days ($24K) is well below the trailing pace — Nvidia’s sponsored YouTube cadence has cooled meaningfully. The $0 in the last 7 days reinforces that (no detected sponsored placements in the most recent week).
- Engagement and views are both negative on the 30-day growth window — fewer placements and softer per-piece performance.
- Instagram is excluded. Nvidia has 0 sponsored creators indexed on Instagram, and
/sponsor/summary doesn’t model IG spend regardless — so this number is effectively the full picture for the platforms CreatorDB tracks (YouTube + IG), but it does not include TikTok or other channels.
- These are modeled estimates, not booked spend — useful for benchmarking and trend, not for citing as Nvidia’s actual budget.
- Engagement rate here uses (likes + comments) / views, different from creator-side engagement formulas — don’t compare directly to creator search filters.
Credits used: 25 · Remaining: 991,910
Example 6: Transcribe a video for content analysis
To transcribe a video for content analysis, you can use the content-analysis skill. For example, to transcribe a creator’s video to find out if they mention a specific brand, you can use the following prompt:
Prompt example
Response example
Claude Code will then use the content-analysis skill to call the CreatorDB API. MrBeast's Most Recent Transcribable Videos — Brand Mentions
The transcript availability situationOf MrBeast’s most recent uploads:The actual most-recent long-form video still hasn’t had subtitles ingested by CreatorDB.Transcripts of the two most-recent transcribable videospzxtGL9kzPs “Try Not To Get Slimed” (2026-04-24, ~30s)
“OH, GOSH. WHAT IS THAT? WOAH, WOAH, WOAH, WHAT? ARE YOU DONE? YEAH, GO AWAY. THAT’S DISGUSTING. … OH, IT FEELS SO NASTY, BRO. KEEP POURING. ALL RIGHT, HERE WE GO. … IT’S SO BUBBLY.”
Brand mentions: none. Stunt reaction video, no dialogue beyond exclamations.yy-Nt0dRhNI “Winner Picks Our Honeymoon” (2026-04-21, ~45s)
“Winner picks our honeymoon. Okay, you just got to get more balls in than the other person … I want to go on an island. … Right after we get married WITHOUT MY PHONE, BABE. … I want uh to go to space. … Hey, Jeff Bezos, you know that like rocket that close your ears that looks kind of like a […] that goes to space. I would love to go.”
Brand / public-figure mentions
creator-csv-workflow skill. For example, if you have a CSV file with a list of YouTube creator channel IDs and you want to enrich it with additional data from CreatorDB, you can follow these steps:
Get a list of creator account or channel IDs.
a. Obtain YouTube creator channel IDs through any of the following methods:
You can call the CreatorDB API using the creator-search skill in Claude Code to find creators based on specific criteria (e.g., niche, subscriber count, location). This will return a list of creator profiles, including their channel IDs, which you can then export to a CSV file for enrichment. For details, see Examples 2 and 3 in this guide on how to use the creator-search skill effectively.
- Search social media platforms, such as YouTube.
- Use your own database.
- Purchase a list from a data provider.
b. Ensure that the channel IDs are in a CSV format with a header row (e.g., channelId).
Prompt Claude Code to enrich the creator list.
Prompt Claude Code with the following query, replacing path/to/your/creators.csv with the actual path to your CSV file:Claude Code will then use the creator-csv-workflow and creator-enrichment skills to call the CreatorDB API and enrich your creator list with additional data such as subscriber count, engagement metrics, audience demographics, and more. The final output should be an enriched CSV file with the new data columns added, which you can use for further analysis or outreach.
For example, if your original CSV had a channelId column, the enriched CSV might include additional columns like displayName, totalSubscribers, engagementRate, audienceAgeRange, etc., depending on the data available for each creator in CreatorDB.
The following is an example of the messages returned for a successful enrichment of a list of 50 gaming YouTubers:ResultsBio-first saved ~165 cr vs always calling /contact (would’ve been 50 × 27 = 1,350).Output
- Enriched data:
/{YOUR WORKING DIRECTORY}/{FILENAME}.csv — 50 rows × 40 columns (profile + audience demographics + emails + _row_id / email_source / _credits_used / _error bookkeeping).
- Summary:
/{YOUR WORKING DIRECTORY}/{FILENAME}.csv — totals + failed-row list (empty).
- Script:
/{YOUR WORKING DIRECTORY}/{SCRIPTNAME}.py (kept, in case you want to re-run or adapt).
13 channels with no email found. Next steps and resources
For further details about the CreatorDB API endpoints used in these examples and costs, please refer to the following documentation:
