Data infrastructure
for the creator economy

InfluencerUnion is a continuously-updated public database of creators across YouTube today, with TikTok and Instagram in the pipeline. Every profile carries the underlying audience metrics from the platform's own public APIs, a niche label produced by a transparent classifier, and an algorithmically-estimated sponsorship rate range with the formula published in full on the methodology page.

The product surface is built for three audiences at once: brands sourcing partnerships, creators benchmarking themselves against peers, and analysts studying the creator economy as a sector. All three see the same data, the same prices, and the same sources.

We are not an agency. We do not book deals on behalf of brands or creators, do not take commissions on partnerships, and do not represent talent.

We are not a marketing platform. We do not run influencer campaigns, host creator-brand workflows, or sell “growth services.” The product is a database and a pricing model, not a managed service layered on top.

We are not a paywalled directory. Every creator page, niche page, country page, and the full methodology are public and free to read without an account. There is no “premium tier” that hides the data behind a credit card.

We are not claiming our pricing estimates are quotes. They are statistical estimates derived from audience size, a niche-specific CPM, and platform / format multipliers. Actual negotiated rates vary substantially. The model and every constant in it are documented so the estimate can be argued with rather than taken on faith.

• Brands sourcing partnerships — from a DTC running their first influencer test to a Fortune 500 comparing 50 creators across regions. The database is comprehensive enough to be a single source of truth and the pricing model is transparent enough to defend internally.
• Creators — benchmark your audience and estimated rate against peers in the same niche and country, and calibrate the model with anonymous deal context when you have it.
• Public-market and IR teams — equity research on creator-economy companies (YouTube, Meta, ByteDance, Roblox, Shopify, on-ramps and tooling) often gets stuck on “how many creators actually monetize, at what scale, in which verticals?” The index is a structured answer to that question, refreshed quarterly with public reports.
• Researchers and journalists — a public data layer with a documented methodology is more citable than scraped tables on a personal blog.

Bottom-up discovery from public data. The index is built from the platforms' own public APIs — YouTube Data API v3 today, with TikTok and Instagram next. We do not buy creator lists from data brokers. Every channel was surfaced by one of four primitives: top-channel seed lists, trending sweeps, niche-targeted search, and comment-thread harvesting. The full discovery pipeline is documented in the methodology page.

Classifier-driven niche labelling. Each creator's niche is decided by a language-model classifier with a confidence threshold. Channels below the threshold sit in a holding bucket rather than being mis-labelled. The classifier sees the channel name, bio, country, language, subscriber count, and a sample of recent video titles — not a black-box embedding of unknown provenance.

Transparent algorithmic pricing. Each rate estimate is audience size × niche CPM × platform multiplier × format multiplier. Every constant is published. Every multiplier is named. There is no proprietary scoring layer on top.

Community calibration. When a creator or brand shares anonymous deal context, that data point feeds the calibration layer above the algorithmic estimate. Individual deals are not displayed publicly — only the aggregate calibration effect on the model. This protects deal confidentiality while still letting the estimate improve.

• Comprehensive over selective. The index tries to include every working creator above a niche-specific subscriber threshold, not just a curated shortlist. Long-tail coverage is the point.
• Transparent over black-box. The pricing formula, every constant, the classifier prompt, the discovery sources, and the limitations are all published on the methodology page. The argument is “here is what we computed and why” rather than “trust the score.”
• Open over paywalled. All creator, niche, country, and cross pages are public and indexable by search engines. We do not gate the data behind a credit-card wall. The eventual paid product is an API and richer historical access — not access to the data itself.
• Honest about limitations. The methodology page ends with a list of known weaknesses in the current model. That list grows, not shrinks, as we learn more.

• TikTok and Instagram pipelines. The YouTube layer is the foundation; the next two platforms extend coverage to the formats where most non-long-form creator spend lives today.
• Public API. A documented REST API for programmatic access to the index, with API-key auth and a free tier for individual research use. The endpoints are already running internally — see the API docs.
• Quarterly creator-economy reports. Each quarter we'll publish a downloadable report summarising index-level shifts: niche growth, country shifts, CPM changes by sector, calibration updates. Designed for the IR/research audience that wants citeable numbers, not blog speculation.
• Creator-claimed profiles. Creators can claim their own page and add self-reported deal context that flows through the calibration layer the same way brand submissions do. Claiming does not let a creator edit the audience metrics — those stay sourced from the platform.