1,000,000+ Instagram and TikTok creators behind one REST endpoint.Filter by followers, engagement, niche, location and email.Hire on the same key.
base: https://api.ugcroster.com/v1 · auth: Bearer rsk_…
1M+
Creators in the directory
14
Search filters
15
Endpoint groups
37
MCP tools
Every creator is one structured record.
16 snake_case fields, the same shape every time. We handle the scraping and upkeep.
01
16 fields per record
Handle, followers, engagement rate, niches, bio, location, demographics, plus an extracted email where we have one.
02
14 search filters
Filter on any field. Combine freely, sort, paginate to 100 per page.
03
10 credits per search, 1 per read
Search hits the whole index; reads are cheap. Brand-key writes are never metered.
Two data endpoints. Thirteen more when you hire.
Data keys cover search and profiles. Brand keys add the whole campaign workflow, same base URL.
One Bearer token, JSON in and out. Dashboard, pipeline or agent.
Creator discovery inside your product, without running a scraper.
Free-text q plus 13 structured filters, sorted and paginated.
Followers, engagement, niche, location, demographics, email.
Brand keys add the write side: brief, contract, message, pay.
POST /v1/briefs and /v1/contracts on the same base URL.
Hosted Streamable HTTP or local stdio. No integration code.
Point Claude or Cursor at api.ugcroster.com/v1/mcp with your key.
Every endpoint, exposed as an MCP server.
Claude, Cursor and any MCP client get the full surface with zero integration code: hosted with just your key, or local over stdio.
remote: https://api.ugcroster.com/v1/mcp · local: npx ugcroster-mcp
claude_desktop_config.json
{
"mcpServers": {
"roster": {
"command": "npx",
"args": ["ugcroster-mcp"],
"env": { "ROSTER_API_KEY": "rsk_..." }
}
}
}Where this fits among the alternatives.
Named head-to-heads live on the blog.
What they do
Roster API
Proxy networks & scrapers
Arbitrary pages. You run the proxies, parsers and retries.
Creator records only: parsed, indexed, one stable schema.
Scraper marketplaces
A scraper rented per job, output shaped by whoever wrote it.
One integration, flat monthly credits.
Influencer analytics suites
Analysts in a dashboard. Seat-priced; API sold as an add-on.
Engineers. The API is the product.
Creator connect-flow APIs
Creators who already authorised your app.
Creators you have not met yet: top-of-funnel search.
Category shapes, not named-vendor claims.
Proxy networks & scrapers
Arbitrary pages. You run the proxies, parsers and retries.
Roster API
Creator records only: parsed, indexed, one stable schema.
Scraper marketplaces
A scraper rented per job, output shaped by whoever wrote it.
Roster API
One integration, flat monthly credits.
Influencer analytics suites
Analysts in a dashboard. Seat-priced; API sold as an add-on.
Roster API
Engineers. The API is the product.
Creator connect-flow APIs
Creators who already authorised your app.
Roster API
Creators you have not met yet: top-of-funnel search.
Pay for what you pull.
Searches cost 10 credits, reads cost 1. Tiers differ only on volume and rate.
One product, real traffic.
25k credits / month
- 25,000 credits / month
- 2,500 searches or 25,000 profile reads
- 14 search filters
- 60 requests / minute
- Email support
Discovery at volume.
150k credits / month
- 150,000 credits / month
- 15,000 searches or 150,000 profile reads
- 14 search filters
- 120 requests / minute
- Priority support
You are the data layer for someone else.
500k credits / month
- 500,000 credits / month
- 50,000 searches or 500,000 profile reads
- 300 requests / minute
- Usage reporting
- Dedicated support channel
On a Roster brand plan? Your brand key covers the full API, unmetered. Manage it in Settings → API & Integrations.
Notes from the team running the index.
Integration guides, comparisons and data deep-dives.
Straight answers.
Structured creator records from a directory of 1,000,000+ Instagram and TikTok creators: username, full name, follower count, engagement rate (a 0-1 ratio), category, niches, biography, city/state/country, age, gender, verification status and, where available, an extracted contact email. Search supports 14 filters and returns up to 100 records per page.
A scraper is infrastructure you operate: proxies, bans, retries, and a parser that breaks when a layout ships. This is a maintained index behind a REST endpoint: one stable schema, one Bearer token, no infrastructure on your side. If you need arbitrary pages from anywhere on the web, use a scraper. If you need searchable creator records, use this.
Credits. A search costs 10 credits, a profile read costs 1. Starter is $49/month for 25,000 credits, Growth is $199/month for 150,000, Scale is $499/month for 500,000. There is no free tier; there is no per-seat pricing. Need more volume? Talk to the team.
60 requests/minute on Starter, 120 on Growth, 300 on Scale. Exceeding the per-minute limit returns 429 with code RATE_LIMITED. Exhausting the monthly credit budget returns 402 with code CREDITS_EXHAUSTED.
A data key is what you buy here: self-serve, read-only access to creator search and creator profiles, metered in credits. A brand key comes free with any Roster brand plan and covers the full surface: briefs, campaigns, applications, contracts, messages, payouts, shipments, webhooks. Same rsk_ format, same base URL.
Yes, and that is the difference from every data vendor. The records come from a working creator marketplace. Find creators on a data key; when you want to brief, contract and pay them programmatically, a Roster brand plan adds the write endpoints on the same API.
https://api.ugcroster.com/v1. Authenticate with Authorization: Bearer rsk_... on every request. Use this host: the apex domain redirects, and most HTTP clients drop the Authorization header on a redirect.
Yes, two ways to connect. Hosted: point Claude, Cursor or any Streamable-HTTP MCP client at https://api.ugcroster.com/v1/mcp with your key, no install. Local: npx ugcroster-mcp over stdio. Either way an agent can search creators and (on a brand key) run campaigns without custom integration code.
That depends on what is collected and how you use it, and you should ask your own counsel. The practical difference: operating your own collector puts those questions on you; licensing an API moves the collection relationship to the provider.
Grab a key, paste one curl, and you're querying the directory. Keys are self-serve.