HumanHours icon

HumanHours

Track AI agent tasks and enrich company domains into a labour-cost business case

Actions7

Event → Track Event

AI-generated

Summary

Track a single AI agent task event by sending details such as agent ID, task type, outcome, and optional additional metrics to the humanhours API.

Inputs

  • Agent ID (required) — Slug identifying the agent that performed the task.
  • Task Type (required) — Slug of the task type. Use either a built-in humanhours key or a custom key.
  • Outcome (required) — Task outcome, chosen from Failure, Needs Review, or Success.
  • Additional Fields — Optional fields including agent duration (seconds), human baseline (minutes), metadata (JSON object), model identifier, tokens in, and tokens out to provide richer event data and assist in cost/pricing calculations.

Output shape

a list of JSON objects representing the tracked event responses from the humanhours API, one per input item.

Each output item is paired with its corresponding input item. The node handles multiple input items by sending individual track requests for each. Errors in individual items can be optionally caught if 'Continue On Fail' is enabled.

Examples

Example 1: Track an email classification task completed successfully by an agent with timing and model token details.

Provide the agent ID slug, task type slug "email_classification", select outcome "Success", and include additional fields for duration, model, and token counts.

Links

Discussion