Data and retention

Everything this API returns is either public short-form video data or something computed from it. This page is a developer’s summary of the Terms and the Privacy Policy: what we process, on what basis, how long we keep it, and what you may build with the results. Those two pages are the authority. Where this summary is shorter than they are, read them.

What we process

The API analyses short-form video published publicly on TikTok, Instagram and YouTube. To do that we collect and process material that is already public on those platforms:

• Public posts and their captions, hashtags, sounds, thumbnails and publicly reported metrics such as view, like, comment and share counts.
• Public creator profile information: handle, display name, biography, and follower and post counts.
• Public comment text on those posts, and derived analysis of it.
• Public advertising records made available through the Meta and TikTok ad libraries.

We obtain most of this through third-party data providers, including ScrapeCreators, and from the platforms’ own public surfaces and advertising libraries.

On what basis

Some of this material relates to identifiable people — creators, and the authors of public comments. Where it is personal data we process it on the basis of legitimate interests (Article 6(1)(f) UK/EU GDPR and equivalent provisions elsewhere): operating a research and analytics service that helps customers understand what is performing on public short-form video platforms. That interest has been assessed against the rights of the people concerned, and the processing is limited accordingly:

• Only material the platform has made publicly accessible. No private accounts, restricted posts, direct messages or anything behind a login that is not our own.
• The data describes content and public performance. It is not used to build profiles of individuals, and not for advertising, credit, employment, insurance or any other decision about a person.
• Comment analysis is stored without usernames or author identifiers, so what is retained describes the response to a video rather than who said what.
• Special-category data is not deliberately collected, and the Terms prohibit customers from using the output to identify, target, profile or make decisions about individuals.

You can see that last rule in the shape of the API itself. GET /v1/videos/{platform}/{id}/comments returns comment text with no author field of any kind — it is stripped at the vendor boundary, before it reaches us, so there is nothing to redact later.

{
  "run": {
    "id": "run_8f2c1ad04b",
    "kind": "video_comments",
    "status": "completed",
    "created_at": "2026-09-08T09:14:02.418Z",
    "completed_at": "2026-09-08T09:14:04.930Z",
    "charge_usd": 0.06,
    "list_price_usd": 0.06,
    "cached": false
  },
  "result": {
    "platform": "tiktok",
    "video_id": "7300000000000000000",
    "items": [
      {
        "id": "7300100000000000001",
        "text": "Does the temperature matter more than the time?",
        "like_count": 4821,
        "reply_count": 12,
        "posted_at": "2026-09-06T08:41:00.000Z",
        "parent_id": null,
        "fetched_at": "2026-09-08T09:14:04.219Z"
      },
      {
        "id": "7300100000000000002",
        "text": "started last week and the first minute is still brutal",
        "like_count": 218,
        "reply_count": 0,
        "posted_at": "2026-09-06T09:02:00.000Z",
        "parent_id": "7300100000000000001",
        "fetched_at": "2026-09-08T09:14:04.219Z"
      }
    ],
    "next_cursor": "eyJvIjoyNX0",
    "analysis": {
      "version": "v1",
      "sample_size": 25,
      "sentiment": {
        "positive": 16,
        "neutral": 6,
        "negative": 3
      },
      "average_score": 0.41,
      "themes": [
        {
          "label": "water temperature",
          "mentions": 9,
          "sentiment": "neutral"
        },
        {
          "label": "cost of a tub",
          "mentions": 4,
          "sentiment": "negative"
        }
      ],
      "computed_at": "2026-09-08T09:14:04.780Z"
    }
  }
}

{
  "estimate_usd": 0.06,
  "cached": false
}

How long we keep it

Records and Snapshots

Video and creator records, and the Snapshots of public metrics taken over time, are retained indefinitely. The historical series is what makes change measurable; deleting it would remove the core function.

Video media

Media files are deleted once the analysis pass over them is complete. We do not keep a library of downloaded videos.

Raw comments

Raw comment text is deleted after 90 days. The derived analysis is kept, and never carries usernames.

The indefinite half is what you are buying: GET /v1/creators/{platform}/{handle}/snapshots reaches back before your first call, because the series belongs to the subject rather than to the Customer who first asked about it. The other two halves are why a 90-day-old video’s comment analysis is still readable when its raw comments are not, and why the analysis of a video outlives the media it was computed from.

What you may do with results

The API returns two kinds of material: raw records — public posts, captions, public metrics, public profile fields and comment text as published on TikTok, Instagram and YouTube — and derived output, which is what we compute from those records: normalised structures, analysis objects, baselines, Scores, Snapshots, Change Reports and Digests.

Derived works are permitted.

• Analyse, transform and combine the API output, and build on it.
• Distribute and commercialise what you build — reports, dashboards, scores, recommendations, editorial analysis — including to your own customers and to the end users of your agent or application. You own what you create this way.
• Display individual raw records incidentally in your own product: quoting a caption, showing a thumbnail, citing a metric to support an analysis. That is not bulk republication.

Reselling the raw data is not. You may not:

• Reselling, sublicensing or redistributing the raw records or bulk API output as a data product, feed, export or dataset, whether for a fee or free of charge.
• Republishing raw records in bulk, or mirroring, replicating or reconstructing our stored data or any substantial part of it, including by systematic enumeration of the API.
• Using the API or its output to build, train or operate a substitute for the Service, or a competing short-form video data service.
• Using the output to identify, target, profile, score or make decisions about individual people, including creators and commenters, or enabling anyone else to do so.
• Removing, obscuring or misrepresenting the provenance of data obtained through the Service where you are permitted to display it.

Upstream restrictions pass through. Some raw data reaches us through third-party providers, including ScrapeCreators, and through the platforms’ own surfaces. Where those providers or platforms restrict the use, retention or redistribution of that data, those restrictions apply to you as well. If an upstream restriction and the Terms conflict, the more restrictive applies.

Accuracy

The Service reports public data and computations over it. Metrics change, platforms remove content, and enrichment is produced by AI models. Output may be incomplete or wrong, and coverage for some platforms is explicitly marked as partial. You are responsible for verifying anything you rely on.

Getting in touch

Questions about privacy, and requests from creators or commenters whose public content we process — including objecting to processing carried out on the basis of legitimate interests, and access, rectification and erasure — go to privacy@shortsintel.com. Include enough detail to identify the account or content in question, such as the platform and handle. We respond within the period the applicable law requires, and you may also complain to your local data-protection authority.

Questions about the Terms, including whether something you want to build counts as a derived work, go to legal@shortsintel.com. Ask before you ship rather than after.

Full text: Terms and Conditions and Privacy Policy.