Streamforge maintains its own data on 48M creators across YouTube, Twitch, TikTok, Instagram, and X. We collect it ourselves, link each creator’s profiles across platforms when the evidence supports a match, and use AI to analyze both the creator and the audience around them.
Collecting a profile is the easy part. What makes the data useful is what happens afterwards, and that is what this page covers: where the data comes from, how it becomes a comparable creator profile, how the audience method was validated, how IGDB-backed game context is attached, and how all of it has been tested.
Why does Streamforge collect its own data?
Because the numbers we were being shown didn’t hold up. Streamforge started in 2018 as a Discord community of gaming creators, and running it meant living inside Twitch’s API every day. When we compared what influencer marketing platforms were reporting against what the platform data actually said, the gap wasn’t subtle. We decided the only data we could trust was data we collected ourselves.
We then ran creator campaigns for gaming publishers on that database for years, and every claim the data made got tested against a real campaign. The platform opened to customers in March 2023. The database came first, and the product was built on top of it.
What does “48M creators” mean?
It is the number of creators in Streamforge’s database. We publish it as a fixed figure confirmed against the current database count, and update it when the database is measured again, instead of showing a counter that changes every time the site loads.
The figure describes the creator data itself: profiles, content, and performance. Analysis is built on top of that. In the dashboard, audience analysis is generated when a profile is opened, and optional fields such as contact details depend on the evidence available for that creator.
Which creator platforms does Streamforge cover?
YouTube, Twitch, TikTok, Instagram, and X. Each platform has its own content formats, account structures, and metrics, and Streamforge keeps those differences instead of flattening them, so a Twitch stream is still read as a stream and a TikTok clip as a clip.
Jynxzi
@jynxzi
- Followers
- 11.6M
Oct 15
MrBeast
@mrbeast
- Followers
- 521M
- Views
- 142.8B
Oct 15
Tommy
@tommyinnit
- Followers
- 11.7M
Oct 15
pokimane
@imane
- Followers
- 5.8M
Oct 15
Mark
@markiplier
- Followers
- 12.7M
- Following
- 379
- Posts
- 9,024
Oct 10
Audience analysis is supported across all five, and the External API covers all five. Individual features and endpoints can differ, and the product and API documentation list what each one supports.
How does Streamforge collect and normalize creator data?
Streamforge uses a five-stage process:
- 01
Platform-specific ingestion
Dedicated services collect creator, content, and performance records in the format supplied by each platform.
- 02
Normalization
Common concepts such as creator identity, reach, content, engagement, and category are mapped into a shared model while platform-specific facts are preserved.
- 03
Identity linking
Profiles are connected across platforms when the available evidence supports a match, so a creator’s channels can be read as one person.
- 04
AI analysis
Streamforge analyzes creator evidence and audience signals to generate qualitative creator intelligence plus demographic and psychographic audience estimates.
- 05
Product availability
The resulting fields reach customers through the dashboard and the API.
This approach gives teams a consistent way to search and compare creators while preserving important differences between a Twitch stream, a YouTube video, an Instagram post, and a TikTok clip.
How does Streamforge add game and genre context?
Streamforge uses IGDB as its source of truth for game and genre metadata. Proprietary entity-resolution techniques connect that taxonomy to creator identities, channels, and content across social platforms.
The system can identify the game associated with an individual stream, video, post, or clip even when the source platform does not provide native game tagging. That makes it possible to distinguish real title and genre relevance from a generic keyword match.
What are creator analysis and audience analysis?
Creator analysis describes the person behind the channel. Streamforge combines creator profiles, published content, and other public sources from around the web. The result can include both directly supported facts and model-inferred qualitative details such as gender, backstory, interests, places visited, relationships, business ventures, content style, and how the creator presents themselves.
Audience analysis describes the people who follow or interact with the creator. Streamforge’s proprietary AI processes comment data and available profile data from those social-media users to generate demographic and psychographic estimates. It does not require the creator to authenticate through OAuth, and the model was trained on first-party audience data.
| Comparison | Creator analysis | Audience analysis |
|---|---|---|
| Main question | Who is the person behind the channel? | Who follows or interacts with this creator? |
| Typical evidence | Profiles, published content, and other public web sources | Comments and available profiles from followers and interacting users |
| Example outputs | Gender, backstory, interests, places visited, relationships, ventures, content style, and safety context | Demographics, psychographics, interests, affinities, engagement, community health, and authenticity |
| Coverage | Built from the public evidence available for each creator | All five platforms, with no creator login required |
Creator analysis can combine source-backed observations and inference. Audience outputs are modeled from audience signals, not copied from a creator’s private analytics. One explains the creator; the other explains the audience around them.
Which creators can be analyzed?
Audience analysis is supported across YouTube, Twitch, TikTok, Instagram, and X, and it is not limited to creators who have opted in or connected an account. In the dashboard, the analysis is generated when a profile is opened.
Both kinds of analysis work from evidence. A creator with little public content, or an audience that rarely comments, gives the models less to read. When there is not enough to support a field, it should read not enough data rather than show a guess, so an empty field is never mistaken for a negative result.
How often is Streamforge creator data updated?
Each field is refreshed on its own schedule. Refresh timing varies by:
- Platform
- Creator size and activity
- Data type
- Source availability
- The cost and rate limits associated with collecting the field
Profile statistics, descriptive metadata, content records, and AI-generated analysis can therefore carry different update times. The useful question is which field was updated, for which platform, and when, and that is how Streamforge treats freshness: field by field, not as one label for an entire creator profile.
How is the data kept accurate day to day?
Creator numbers move every day, and a database this size goes stale fast if nobody tends it. Most of that upkeep runs on its own:
Each creator is on their own schedule
Every creator has their own refresh schedule. How often depends on their size and, on most platforms, how actively they post. New YouTube and TikTok videos keep being re-checked in their first weeks, and Twitch streams are tracked while they are live.
Opening a profile checks for newer data
Opening a YouTube, TikTok, Instagram or X creator in the dashboard checks for newer data on the spot.
Bought views are flagged
On TikTok, Instagram and YouTube, Streamforge flags posts whose views were inflated by a brand paying to promote them, and estimates how many views were paid for. The method was checked against campaigns where the real paid and organic split was known.
Comments are checked for bots
Streamforge analyzes the comments on a creator’s posts, videos and streams using AI to detect bot behavior, writing styles, and repetitive phrases or comments that do not engage with the content.
The boost check matters most when you compare creators. A post a brand paid to promote can reach far more people than the creator normally does, and judging the creator on it would overstate what they earn on their own.
How has the data been tested?
In three ways. First, on our own campaigns: before Streamforge was a product, we ran creator campaigns for gaming publishers on this data for years.
Second, against the platforms. Streamforge has benchmarked its audience-analysis technique against first-party audience data from the platforms and found it to be highly accurate.
Third, by buyers. Fortune 500 companies have tested Streamforge’s creator data side by side against other vendors. Streamforge came out ahead in every completed evaluation.
How should teams interpret creator metrics across platforms?
Normalized data makes cross-platform discovery possible, but platform metrics still retain their original meaning. Followers, subscribers, views, live viewers, posts, videos, and engagement actions are not interchangeable.
Teams should compare creators using the metrics that fit the campaign and content format, then use available audience and content evidence to refine the shortlist. A single number rarely tells the whole story.
