Perplexity Decisions API can search the web before it answers, with mode web_grounding (November 2026)

What you get

In November 2026 Perplexity added web grounding to the Decisions API (POST /v1/decisions). Set mode to "web_grounding" and the API runs one Fast Search per question, adds the top results to your state as evidence, then answers. The response gains a search_results list with each result's title, URL, snippet and, where known, dates. Requests without mode work as before.

• Price: $0.001 per search, so at most $0.10 for 100 questions. The evidence is billed as input tokens and counts toward the 256K input limit.
• Limits: up to 100 questions per grounded request (more returns 400). Each search returns up to 5 results and the API keeps up to 10. Searches run up to 10 at a time, 1 second each and 3 seconds in total; a search that times out adds nothing. Images aren't searched.
• For many questions Perplexity recommends pplx-decider-v1.2 over pplx-decider-flash-v1.2, which counts the evidence again for every question.

Prerequisites

• A Perplexity API key in PERPLEXITY_API_KEY
• Any HTTP client, or the Python SDK 0.44.0 or later, or the TypeScript SDK 0.39.0 or later

Steps

1. Add "mode": "web_grounding" to a Decisions request:

curl -X POST https://api.perplexity.ai/v1/decisions \
  -H "Authorization: Bearer $PERPLEXITY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "pplx-decider-latest",
    "mode": "web_grounding",
    "state": "Claim: The James Webb Space Telescope launched in December 2021.",
    "questions": {
      "supported": {
        "type": "noul",
        "instructions": "Is the claim true?"
      }
    }
  }' | jq