Perplexity’s Orchestrator API: One Key for Claude, GPT, Gemini and More

Perplexity's Orchestrator API: One Key for Claude, GPT, Gemini and More

Written by

in

Perplexity is testing the Orchestrator API, which lets developers reach many companies’ AI models with one key, one address and one bill. It is in private preview, and developers have to request access.

Key facts

  • It covers models from Anthropic, OpenAI, Google, xAI and open-weight labs such as DeepSeek, GLM and Kimi, which Perplexity hosts itself.
  • It accepts the request formats developers already use for OpenAI and Anthropic, so most existing code only needs a new address and key.
  • You pay each model’s own per-token rate, with no extra fee per request.
  • The product has changed names twice: it started as the Gateway API in August 2026, became the Router API, and is now the Orchestrator API.

What it means for you

This is mainly for developers. If you or your small business build AI tools, it could make it easier to try several models and keep billing in one place.

Source: Perplexity AI Magazine, Perplexity Orchestrator API Guide. Summary and notes by My AI Lighthouse.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *