@pipedog/perplexity

Perplexity services for pipedog: PerplexityAgentHttpLLMService, an LLM service for Perplexity's Agent API (web search over the OpenAI Responses protocol), and PerplexityLLMAdapter. The service extends OpenAIResponsesHttpLLMService from @pipedog/openai, so streaming, function calling, usage metrics and runInference work as they do there. Upstream's PerplexityLLMService (Sonar chat completions) is deprecated and not ported.

A port of Pipecat's pipecat.services.perplexity and pipecat.adapters.llm.services.perplexity_adapter. Subpaths mirror upstream: @pipedog/perplexity/llm and @pipedog/perplexity/perplexity_adapter.

Install

bun add @pipedog/core @pipedog/asyncio @pipedog/perplexity

It depends on @pipedog/openai (and the official openai JS SDK), which is installed with it.

Usage

import { PerplexityAgentHttpLLMService } from "@pipedog/perplexity/llm";

const llm = new PerplexityAgentHttpLLMService({
  apiKey: process.env.PERPLEXITY_API_KEY ?? "",
  settings: new PerplexityAgentHttpLLMService.Settings({
    systemInstruction: "You are a helpful voice assistant. Keep replies short.",
  }),
});
// new Pipeline([transport.input(), stt, userAggregator, llm, tts, transport.output(), ...])

Runtimes

Bun, Node.js 22+, Deno and Cloudflare Workers: the client talks through fetch.

Links

• Docs: Perplexity
• pipedog README
• License: BSD 2-Clause, a port of Pipecat (Copyright (c) 2024–2026 Daily). See LICENSE.