Task

What a task is

A task is Tray.ai’s unit of usage. Each operation Tray performs on a piece of your data counts as one task:

• the trigger firing
• each connector or service step
• each pass through a loop or a boolean

The shorthand in Tray’s documentation is to think of a task as one step in a workflow. A workflow run is therefore several tasks, not one: a five-step workflow that fires once uses about five tasks, and a loop over 100 records uses a task for each pass.

A task is not an API call. A connector step that makes several API calls to do its job still counts as one task. Retries are not charged.

Tasks are observable. Each step in a run has a trace with timing, inputs and outputs, and a success or failure outcome.

Why it matters in pricing

Tray starts at $50 a month for Tray Helix Builder Edition. Tray iPaaS, where tasks apply, is priced differently: a platform fee for your plan, plus tasks for usage. There are no per-user or per-connection fees, larger volumes get preferential pricing, and a free trial shows how many tasks your workflows use before you buy.

To estimate volume, count the steps in each workflow, multiply by how often it runs, and allow for loops over records. See the pricing page for the full structure.

Scale context

Tasks at Tray.ai scale:

• Cisco runs 600M+ tasks per year across 60+ automations.
• DocuSign runs 150M tasks per month.
• Auctane processes 3B orders per year through Tray.ai.

Those numbers give a sense of what “tasks” actually means in production — not a rate limit to worry about, but a normal unit of work at enterprise scale.

Where you see tasks

The Intelligent iPaaS pillar surfaces task counts, per-workflow volume, per-step execution details — the observability layer of the platform. Anywhere you’re running process automation or data integration, tasks are the thing being counted.