The Perplexity AI Connector is a service module built for OutSystems 11, enabling seamless integration with Perplexity AI’s Sonar model. It allows developers to bring cutting-edge conversational intelligence into their applications by sending natural language prompts and retrieving AI-generated responses enriched with web citations, related questions, and contextual images.

This connector can be consumed in both Reactive Web and Mobile apps, making it ideal for scalable, cross-platform intelligent experiences.

What It Does:

• Sends structured natural language prompts to Perplexity AI via its official API.

• Allows control over reasoning settings, tone, and precision using parameters like:

  • search_mode = web, academic

  • reasoning_effort = low, medium, high

  • temperature, top_p, presence_penalty, frequency_penalty

  • Domain-based search filters (e.g., nasa.gov, wikipedia.org)

  • Recency filtering (e.g., week or day)

• Returns:

  • AI-generated answer content

  • Citations (source URLs with metadata)

  • Related follow-up questions

  • Image suggestions for visual enhancement

  • Search result context (title, date, source)

Included Elements:

• Structures:

  • PerplexityRequest – input structure with full control over prompt and parameters

  • PerplexityResponse – output structure capturing answer, images, citations, related questions

• Server Action:

  • SendPromptToPerplexityAI – encapsulates the full API call and response deserialization

• Site Property in Demo:

  • PerplexityAPIKey – securely store and manage your API key

• No UI dependency – can be directly consumed by Reactive or Mobile apps

Use Cases:

• Intelligent chatbots and virtual assistants

• Smart search features within enterprise or consumer apps

• Educational tools or FAQ bots backed by verifiable sources

• Domain-specific Q&A experiences with image and citation support

Requirements:

• A valid Perplexity AI API Key (invite-based access as of now)

This connector delivers a conversational AI experience that is fast, explainable, and rich in context, unlocking new possibilities for AI-driven OutSystems applications.