Enterprise API

PowerTrack

Description:
A highly-reliable, full-fidelity, filtered stream of Tweets delivered in real-time as they are created based on a set of boolean rules.

Product details:
PowerTrack is the enterprise alternative to the standard statuses/filter endpoint with a significantly enhanced filter language and reliability features to avoid data loss.

Common use cases:

  • Topical analysis
  • Event detection
Full-archive Search

Description:
Filtered access to the entire public history of Tweets through boolean queries

Product details:
The enterprise version of the Full-Archive Search API is equivalent to the premium version but supports increased scale and access levels.

Common use cases:

  • Historical topic analysis
  • Historical event analysis
  • Longitudinal studies
30-day Search

Description:
Filtered access to the last 30 days of the public history of Tweets through boolean queries

Product details:
The enterprise version of the 30-Day Search API is equivalent to the premium version but supports increased scale and access levels.

Common use cases:

  • Recent topic analysis
  • Recent event analysis
Historical PowerTrack

Description:
Job-based processing of the entire history of Tweets for file download based on boolean filters.

Product details: 
Intended for well-funded enterprises and institutions only, and purchased as an annual subscription, account holders may run multiple jobs by API each month. Data is only available in JSON format (not CSV/TSV) and requires download of a large number of data files in parallel.

Common use cases:

  • Historical topic analysis
  • Historical event analysis
  • Longitudinal studies
Decahose

Description:
A statistically random 10% sample of all Tweets created on the Twitter platform delivered in real-time as they are created

Product details:
The Decahose is the enterprise equivalent of the statuses/sample endpoint.

Common use cases:

  • Trend and event detection
  • Data processing and pipelines
  • Projects related to scale
  • Projects only requiring a sampled dataset