Skip to main content
For AI agents: a documentation index is available at https://docs.parallel.ai/llms.txt. The full text of all docs is at https://docs.parallel.ai/llms-full.txt. You may also fetch any page as Markdown by appending .md to its URL or sending Accept: text/markdown.
Structured outputs enable you to define a JSON schema for monitor events. Each detected event conforms to the specified schema, returning data in a consistent, machine-readable format suitable for downstream processing in databases, analytics pipelines, or automation workflows.
Schema Complexity: Output schemas are currently limited to the complexity supported by the core processor. Use flat schemas with a small number of clearly defined fields.

Defining an Output Schema

Include an output_schema field when creating a monitor:
Response:

Retrieving Structured Events

Events from monitors configured with structured outputs include a result field containing the parsed JSON object:
Response:

Best Practices

  • Include property descriptions: Provide clear description fields for each property to improve extraction accuracy
  • Use primitive types: Limit properties to string and enum types for reliable parsing
  • Maintain flat schemas: Use 3–5 properties with a single-level object structure
  • Define edge case handling: Specify how missing or inapplicable values should be represented
  • Quickstart: Monitor API setup and configuration
  • Events: Event model and event groups
  • Webhooks: Real-time notification delivery