Skip to main content
Date: February 5, 2026 We’re excited to announce Exa Instant, an improved neural search with both quality and latency benefits, designed for real-time applications where speed is critical.
Try Instant Search in our API with type = "instant". Try Instant Search in the dashboard

​
What’s New

Sub-150ms latency: Exa Instant delivers our fastest search response times, optimized for applications where every millisecond counts. Real-time use cases: Perfect for low-latency products like chat apps and voice AI, coding agents that need fast web lookups, autocomplete, and live suggestions. Enhanced neural search: Exa Instant combines our best neural search technology with optimized infrastructure to deliver state-of-the-art quality at unprecedented speed. Use type="instant" in your search requests: 
curl -X POST https://api.exa.ai/search \
  -H "x-api-key: EXA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "What is the capital of France?",
    "type": "instant",
    "numResults": 10,
    "contents": {
      "highlights": {
        "maxCharacters": 1000
      }
    }
  }'

​
Python SDK

from exa_py import Exa

exa = Exa('YOUR_EXA_API_KEY')

results = exa.search(
    "What is the capital of France?",
    type="instant",
    num_results=10,
    contents = {
      "highlights": {
        "max_characters": 1000
      }
    }
)

print(results)

​
TypeScript SDK

import Exa from 'exa-js';

const exa = new Exa('YOUR_EXA_API_KEY');

const results = await exa.search(
    'What is the capital of France?',
    {
        type: 'instant',
        numResults: 10,
        contents: {
          highlights: {
            maxCharacters: 1000
          }
        }
    }
);

console.log(results);

​
Search Type Comparison

TypeLatencyBest For
auto~1sHighest quality results (default)
instantsub-150msReal-time applications, autocomplete
fast~500msBalance of speed and quality
deep~5sComprehensive research tasks

​
Need Help?

If you have questions about Instant Search or want to learn more about optimizing for latency, reach out to hello@exa.ai.