Skip to main content
Dust Apps are deprecated. Only Dust Apps created before October 2025 are accessible. The creation of new Dust Apps is deactivated. For building on top of Dust, see the Developer Platform.
A Dust App is a chain of one or more prompted calls to models or external services (such as APIs or Data Sources) designed to perform a particular task. A Dust App is an orchestration layer that sits on top of a model to specialize its behavior for a specific task.

​
Core idea

The design of a Dust App is the process of defining prompts and chains of large language model calls or external service calls to expose a structured API with predictable behavior. Dust Apps are composed of Blocks executed sequentially. Each block produces outputs and can reference the outputs of previously executed blocks.

​
Example apps

​
Translation (Easy)

Translation can be encoded using a few-shot prompt. A Dust App for translation consists of:
  • An input block to receive the text to translate
  • A data block that returns few-shot examples from a dataset
  • An llm block that parametrizes the call to the model
  • A code block to process the model output

​
Recursive Summarization (Medium)

Recursive summarization handles text that is too long for a single model context. The technique splits the text into chunks, summarizes each separately, then concatenates and produces a final summary.

​
Multi-step Research (Advanced)

Multi-step apps chain several model calls, data source queries, and code blocks together to perform complex research or reasoning tasks.