Perplexity API

Perplexity provides an API and developer SDK that combine web search results with large language model reasoning to produce grounded answers and research outputs. The platform exposes a Search API billed per request and a set of Sonar models billed by tokens, including specialized Deep Research variants that return citation and reasoning tokens. Perplexity documents OpenAI-compatible chat completions endpoints and developer tooling (SDK, quickstarts, and examples) to integrate grounded LLMs into applications.

• Search API (per-request billing) — Use the Search API to retrieve raw web search results and filters; start by calling the Search endpoint and supply query and filter parameters. Pricing is request-based.
• Sonar models (token pricing) — The Sonar family (Sonar, Sonar Pro, Sonar Reasoning, Sonar Deep Research) charges by input/output/citation tokens; include tokens in your usage calculations and set model parameters in the SDK or API calls.
• OpenAI-compatible chat completions — The API supports a chat-completions style interface compatible with OpenAI formats; follow the chat-completions SDK quickstart to migrate or integrate.
• Developer resources and community — The docs provide quickstarts, API reference, a cookbook on GitHub, and a community forum and Discord for support. Start with the Perplexity SDK quickstart and create an API key in the dashboard.
• Billing and admin controls — The docs cover API groups, billing setup, rate limits, and key management so teams can monitor spend and control access.