UiPath Documentation
activities
latest
false
Integration Service activities

Generate Async Chat Completion

Generate Async Chat Completion activity in the Perplexity connector.

Note:

As of September 30, 2026, this activity uses Perplexity's Agent API. Perplexity is retiring the Sonar asynchronous chat API that this activity previously used (supported until September 27, 2026), so UiPath migrated the activity to the Agent API to keep your automations running. Existing workflows keep working: saved Sonar model IDs are mapped to the matching preset automatically. A saved Search mode value of academic or sec now fails the request, because the Agent API has no equivalent.

Description​

Creates a chat completion job asynchronously. A job ID is generated as soon as the prompt is submitted. Use the Get Async Chat Object activity or the Completed Asynchronous Chat Job trigger to retrieve the result once the job finishes.

Project compatibility​

Windows | Cross-platform

Configuration​

  • Connection - The connection established in Integration Service. Access the dropdown menu to choose, add, or manage connections.
  • Model - The preset that selects the model and research depth: fast, low, medium, high, or xhigh. Deeper presets run more search steps and take longer. Legacy Sonar model IDs are mapped to the matching preset automatically, and a full model ID is also accepted. For details, refer to the Perplexity documentation for presets.
  • Role - The role of the prompt message: user (default), system, developer, or assistant. Optional.
  • Prompt - The message to send to the model.

Manage Properties​

Use the Manage Properties wizard to configure or use any of the object's standard or custom fields. You can select fields to add them to the activity canvas. The added standard or custom fields are available in the Properties panel (in Studio Desktop) or under Show additional properties (in Studio Web).

Additional properties​

Options​
  • Instructions - System instructions that steer how the agent approaches the task.
  • Maximum steps - The upper bound on the agent research loop, from 1 to 100. Each step is a search or fetch plus a model turn. Lower values reduce latency and cost. If not set, the preset decides: 1 (fast), 5 (low), 15 (medium and high), or 100 (xhigh).
  • Reasoning effort - How much reasoning the model applies before answering: minimal, low, medium, high, xhigh, or max. Leave it empty to use the preset's own level. Available only for the medium, high, and xhigh presets.
  • Temperature - The sampling temperature, from 0 to 2. Higher values make the answer more varied, lower values more focused. Available only for the fast preset.
  • Top p - The nucleus sampling threshold, from 0 to 1. Available only for the fast preset.
  • Maximum tokens - The maximum number of tokens the model can generate. When the limit is reached, the job ends with Vendor status incomplete.
Output​
  • Async chat completion - Automatically generated output variable.
  • Description​
  • Project compatibility​
  • Configuration​
  • Manage Properties​
  • Additional properties​

Was this page helpful?

Connect

Need help? Support

Want to learn? UiPath Academy

Have questions? UiPath Forum

Stay updated