APITube Help Center

How story clustering and duplicate detection work

Find related coverage and filter repeated articles

Tasha Tatum

Written by Tasha Tatum

July 5, 2026

Updated October 16, 2026

Open this example in the API Playground ↗

How story clustering and duplicate detection work

Use a story to find related coverage from different publishers. Use is_duplicate to include or exclude articles flagged as near-identical reprints. A story helps you compare reporting; the duplicate filter helps reduce repetition.

See what is a story cluster for the difference between one article and a group of related reports.

How does APITube build a story cluster?

Choose an article as your starting point, also called the seed, and pass its id to /v1/news/story/:id:

curl "https://api.apitube.io/v1/news/story/1234567?api_key=YOUR_API_KEY"

The response contains other articles with related headlines, ordered by relevance. It excludes the starting article itself.

You can use any article ID. A missing article returns 404 ER0235; a headline with no usable search terms returns 404 ER0240. The endpoint supports GET and POST, with filters in the URL or a JSON body, and returns paginated results.

How does APITube rank the articles in a story?

More closely matching headlines appear first, followed by looser matches. The relationship is based on headlines, so review results when different events use similar words.

A story is useful for comparing coverage, but does not guarantee that every result describes exactly the same event.

How does duplicate detection work, and how do you filter it?

Each article has an is_duplicate flag. Set the same parameter in a search:

  • is_duplicate=0: exclude articles flagged as duplicates.
  • is_duplicate=1: return only flagged duplicates.
curl "https://api.apitube.io/v1/news/everything?title=election&is_duplicate=0&api_key=YOUR_API_KEY"

Only 0 and 1 are accepted. A non-numeric value returns ER0003; another number returns ER0004.

The flag is approximate. Excluding flagged duplicates can reduce repeated coverage, but does not remove every syndicated copy. See how to exclude paywalled or duplicate articles.

Story clustering vs duplicate detection: what’s the difference?

Stories bring related reporting together, including articles written differently by different sources. Duplicate filtering selects or removes articles marked as copies.

You can use both: find related coverage with the story endpoint, then exclude flagged duplicates where you do not want repeated text.

Common Questions

Does clustering match on the article body or just the headline?

Related results are based on the starting article’s headline. They do not use its full body, so coverage with different headline wording may be missed.

What happens if the seed headline has no usable keywords?

You get 404 ER0240. 404 ER0235 means the article ID was not found; omitting the ID returns 400 ER0231.

How much does a story request cost?

One credit if at least one related article is returned. An empty successful result costs no credits.

Can I reuse a story cluster later?

Save the starting article’s ID and call /v1/news/story/:id again. There is no separate cluster ID to save. Results can change as new coverage becomes available. See how to get all articles covering the same story.


Related Articles