Perplexity
CertifiedTasks that request chat completions from Perplexity models.
Set your Perplexity API key, choose the target model, and pass in messages or prompts to generate responses with usage details you can log or feed into downstream steps.
Perplexity
Tasks that request chat completions from Perplexity models.
Set your Perplexity API key, choose the target model, and pass in messages or prompts to generate responses with usage details you can log or feed into downstream steps.
How to use the Perplexity plugin
Run Perplexity AI chat completions from Kestra flows.
Authentication
Set apiKey (required) to your Perplexity API key. Store secrets in secrets and apply connection properties globally with plugin defaults.
Tasks
ChatCompletion sends a chat conversation to a Perplexity model — set model (required, e.g. sonar or sonar-pro), messages (required, list of objects with type (SYSTEM, ASSISTANT, or USER) and content), and apiKey (required). Tune generation with temperature (default 0.2), topP (default 0.9), topK (default 0), presencePenalty (default 0.0), frequencyPenalty (default 0.0), and maxTokens. Set jsonResponseSchema to constrain the output to a JSON schema. The output includes outputText (the model's reply) and rawResponse (full JSON payload).
