Posts, video and transcripts
The content itself and everything in it: video files, transcripts, on-screen text, places, products and clothing seen, captions and metrics for every captured post.
Custom creator datasets, built to your spec.
Any slice of 10M+ influencers and creators and everything they post, with 2,000+ signals on each, delivered to your warehouse and kept fresh.
You ask for2M short-form videos with transcripts, on-screen text and scene labels, balanced across 40 niches
We scope it
- Scope
- 2M posts across 40 niches
- Platforms
- TikTok, Instagram, YouTube
- History
- Last 24 months
- Media
- Video files and transcripts
- Delivery
- Parquet to S3, one-off
- Signals
- Posts, video and transcriptsContent and topics
You get
one row per post| Creator | Format | Scene | Transcript |
|---|---|---|---|
 Jalen Cruz@jalen.moves | Reel · 0:42 | Shibuya, Tokyo | “Okay, day two in Tokyo and the jet lag is real…” |
 Pixie Lane@pixieplays | Short · 0:58 | Home studio | “Nobody talks about how good the soundtrack is…” |
 Ruth Callan@ruthcooks | Reel · 1:12 | Kitchen | “Fifteen minutes, one pan, and you are done…” |
 Marcus Hale@marcusmakes | Video · 9:40 | Workshop | “Before you buy a table saw, watch this…” |
- creators to build your dataset from
- 10M+
- signals analyzed per creator
- 2,000+
- platforms: Instagram, YouTube, TikTok
- 3
- refresh cadence, one-off to daily
- Any
A scrape gives you a row. We give you the creator.
Videos watched and understood: where they went, what they wore, what they used. Take the analysis, the video files, or both.
What a scrape gives you
5 fields- Username
- jalen.moves
- Followers
- 412,093
- Engagement
- 3.8%
- Category
- Athlete
- Bio
- move different | LA
No idea what he posts about, who watches, or where he's been.
Jalen Cruz
Instagram, TikTok and YouTube, one creatorTalks about
Sprint trainingStreetwear fitsTravel
Audience
Mostly 18–29 on the US West Coast, into track, sneakers and gym culture.
Brand fit and history
Performance apparelSports nutritionWorked with 3 travel brands
0.88LA, location confidence
Realengagement check
212videos analyzed
TOKYO DAY 20:4251.8K
Transcript
0:04 Okay, day two in Tokyo and the jet lag is real…
Detected in the video
- Places
- Shibuya Crossing, Park Hyatt Tokyo
- Travel
- Japan Airlines, LAX to Haneda
- Wearing
- Adidas track jacket, New Balance 990s
- On screen
- Fujifilm X100V, “TOKYO DAY 2”
- Format
- Travel vlog, fit check
Video file included
Define the dataset by what's in the content.
Follower counts and categories are table stakes. Because every post is already analyzed, you can scope a dataset by what creators actually say, show and do, then choose how deep each record goes.
- Every creator who reviewed a standing desk on YouTube this year
- Posts where your product appears on screen, tagged or not
- Creators who stayed at a specific hotel group in the last six months
- Your own list of up to 20M handles, enriched and refreshed daily
Dataset definition
Scoped with youCreators who
posted a video abouttrail runningin the last90 days
with an audience mostly inUS and Canada
and10K to 500K followers
For each, include
- Creator profile
- Audience and brand fit
- Matching posts and transcripts
- Video files
- + Your custom field
Delivered daily to BigQueryScope yours
2,000+ signals. Take the ones you need.
Pick the signal families your use case needs and leave the rest. Every sample ships with a field-by-field data dictionary.
Content and topics
What each creator actually talks about across every post, not a single category tag. Themes, formats, vibe and key facts.
Audience
Who is really watching: age, region, interests and the communities each creator reaches.
Brands and partnerships
Every brand that appears, paid or not: tagged, mentioned, worn or shown on screen, and how it performed.
Performance and authenticity
Engagement and views benchmarked against similar creators, posting cadence, growth and signs of inflated numbers.
Identity and location
One creator across Instagram, YouTube and TikTok, with where they're based and how sure we are.
From brief to your warehouse in four steps.
- Step 1
Tell us the brief
Who you need, which signals, how much history, where it should land. - Step 2
Review a sample
A sample of your dataset and the full data dictionary, before you commit. - Step 3
Agree scope and price
Priced by records, signals, media and refresh cadence. - Step 4
It lands on schedule
One-off, or refreshed daily, weekly or monthly.
Delivered where your data already lives
Amazon S3BigQuerySnowflakeParquet, CSV or JSON Lines
The details
- Coverage
- 10M+ creators across Instagram, YouTube and TikTok, plus their posts and videos.
- Freshness
- One-off or refreshed on your schedule, up to daily. Every record carries when it was last updated.
- Format
- Parquet, CSV or JSON Lines, plus media files. Stable IDs so every delivery joins cleanly.
- Sourcing
- Built from publicly available data, with documented sourcing and licensing terms your legal team can review.
- Pricing
- Scoped to your dataset: records, signals, media and refresh cadence.
Questions about custom datasets
What's in a custom influencer dataset?
Whatever your brief needs. Creator profiles, posts and videos, transcripts, audience, brand partnerships and performance, chosen from 2,000+ signals and cut to the creators, content and history you specify.
Which platforms do you cover?
Instagram, YouTube and TikTok, across 10M+ creators. Where a creator is on more than one platform, their accounts are linked into one record.
Can I get the video files, not just the metadata?
Yes. Datasets can include the video files alongside transcripts, on-screen text and what's detected in each video, such as places, products and clothing. You can also take the analysis without the media.
Can you enrich our own list of influencers?
Yes. Send us your handles, up to 20M, and we match, enrich and keep them refreshed on your schedule.
How is the data sourced?
Datasets are built from publicly available data, with documented sourcing and licensing terms your legal team can review.
How is the data delivered?
As Parquet, CSV or JSON Lines to Amazon S3, Google BigQuery or Snowflake, either one-off or refreshed on a schedule up to daily. Stable IDs keep every delivery joinable.
Can I see a sample before committing?
Yes. Describe the dataset you need and we'll scope it and send a sample with the full data dictionary before you commit.
How much does a custom dataset cost?
Pricing depends on the number of records, the signals and media included, and how often it's refreshed. We quote once the dataset is scoped.
Tell us the dataset you wish you had.
Describe the creators, content and signals you need. We'll scope it and send a sample with the full data dictionary.
Elliot PadfieldCo-founder