Skip to main content
Tell the agent what to do and what to return. A task defines the input it accepts, the output it produces, and the work the agent performs on the source. You write the schema, Deck handles everything else. Every task run returns the same structured output regardless of which source it runs against.

How tasks fit in

Tasks belong to an agent. When you run a task, you provide a connection and the task input defined by its schema. Deck creates a task run that executes the work. Workflows can chain multiple tasks into a single execution.

Creating tasks

Tasks can be created in the Dashboard with prompting or from a template. They can also be created through the API. Once created, you reference them by ID in your API calls.

Read vs. write

Tasks can be read operations (fetching data) or write operations (performing actions).
TypeExamples
ReadFetch account balance, list transactions, download statements
WriteSubmit a form, make a payment, cancel a reservation
The API treats both the same way. The distinction is in what the agent does on the source.

Input and output schemas

Every task defines a contract through JSON Schema:
  • Input schema validates what you send when running the task. Deck rejects requests that don’t match before the agent starts.
  • Output schema defines the structure of the data the agent returns. Regardless of how the source renders its data, Deck returns a consistent shape.
This contract means your integration code works the same way across sources. A “fetch reservations” task returns the same structure whether the source is Hilton, Marriott, or Hyatt.

Task statuses

StatusMeaning
learningThe agent is learning how to perform the task. You can run it, but expect lower success rates.
testThe agent understands the task but is still improving. Higher success rates than learning.
liveProduction-ready. The agent fully understands the task.
If you update a task, it may move back to learning or test while the agent adapts to the changes.

Deep dives