Tasks & Scheduling

Add tasks in Orchestrator and see them on RAPBot Agent

Tasks are built in Bot Studio, published to Orchestrator, assigned to an agent, and then appear on RAPBot Agent. Scheduling and per-agent configuration are done in Orchestrator.

End-to-End Flow

Bot Studio  →  Task Library  →  Assign to agent  →  RAPBot Agent (within 30 sec)

1. Publish Tasks to Orchestrator

From Studio, use Send Task to Orchestrator on a local task. The task appears in the organization Task Library.

See Bot Studio — local task options and Orchestrator Tasks.

2. Add a Task to an Agent

  1. In Orchestrator, open the agent detail page
  2. In the Tasks section, click the + button to add a task
Orchestrator agent detail page Tasks section with add task button
  1. The Task Library modal opens — select a task and version
  2. Set manual hours, email configuration if needed, and click Import Task
Orchestrator Task Library modal with Import Task button

3. Configure Task Variables

On the agent task list, open Configure for a task to set variables from the Studio start block (payload, topic, file paths, amounts, etc.):

Orchestrator Task Config modal with payload, topic, static_file_path, and amt fields

Click Update to save. Configuration is synced to the agent along with the task package.

4. Tasks Appear on the Agent

Assigned tasks and updated configuration sync to RAPBot Agent automatically within 30 seconds, or immediately when you click Refresh on the Agent Tasks screen.

RAPBot Agent Tasks dashboard showing multiple synced task cards

On the agent, use the task menu to view Config and Schedule synced from Orchestrator. See Agent Interface.

5. Schedule Tasks

Schedules are created in Orchestrator on the agent's task list (schedule icon / modal), not in RAPBot Agent.

SettingDescription
Execution typeScheduled or Triggered (after another task completes)
RecurrenceDaily, weekly, monthly, or specific dates
Time periodStart and end dates
Execution timeOnce or multiple times per day

Schedules use the agent's time zone from when the agent was created.

Full reference: Scheduling Tasks

View schedule on the agent

After sync, open Schedule from the task ⋮ menu on the agent to read the active schedule:

RAPBot Agent Schedule Info dialog showing recurrence and execution time

Manual Runs and Trigger Task API

From Orchestrator you can Run or Stop tasks on a connected agent from the agent task list.

Trigger Task API (task queue)

This is not the same as Orchestrator data queues or the Studio read queue block (which reads transaction items from a data queue).

The Trigger Task API queues a task run on a connected agent. Use it when an external system or another automation needs to start a specific task on demand.

API

MethodPOST
URL{orchestratorURL}/api/triggerTask/agent/{agentId}/task/{taskId}/start
Headerx-api-key: {your api key}
BodyNone

Get the sample URL and IDs from the Trigger Task API option on a task in the Orchestrator agent detail page. API keys are managed under API Keys.

Orchestrator responds with a message such as Triggered Successfully. The task is set to run in next few mins when the agent is deployed and connected.

How it reaches the agent

  1. The API call registers a pending trigger in Orchestrator.
  2. On the next agent sync cycle (within 30 seconds, or immediately after Refresh), Orchestrator sends the trigger to the agent.
  3. The agent writes the request to its local task queue (agent database).
  4. When no other task is running, the agent picks up the queued run and starts execution.

The same API is used by the Studio task queue block in automation flows — it calls the Trigger Task API with Domain URL, Agent ID, Task ID, and API Key.

Queue behavior

SituationBehavior
Agent already running a taskNew trigger waits in the local queue (ready_to_run) until the current run finishes
Another trigger arrives while one is queuedAdditional runs are queued; unattended agents may skip duplicates already pending
Agent not connectedAPI returns an error (agent not deployed or inactive)

Run from Orchestrator UI

Clicking Run on a task in Orchestrator uses the same trigger mechanism — the request is delivered to the agent on the next sync and added to the local task queue before execution.