Skip to main content

Get your Exa API key

​
Overview

The Context API (also called Exa Code) is a powerful tool for coding agents that need fast, efficient web context. It searches over billions of GitHub repos, docs pages, Stack Overflow posts, and more to find the perfect, token-efficient context that agents need to code correctly. This endpoint helps eliminate hallucinations in coding agents by providing real, working code examples from the open source community.

​
Example Use Cases

The Context API excels at finding practical code examples for:
  • Framework usage: “use Exa search in python and make sure content is always livecrawled”
  • API syntax: “use correct syntax for vercel ai sdk to call gpt-5 nano asking it how are you”
  • Development setup: “how to set up a reproducible Nix Rust development environment”
  • Library implementation: “React hooks for state management examples”
  • Best practices: “authentication patterns in NextJS applications”
Basic Code Search 
curl -X POST 'https://api.exa.ai/context' \
  -H 'x-api-key: YOUR-EXA-API-KEY' \
  -H 'Content-Type: application/json' \
  -d '{
    "query": "how to use React hooks for state management",
    "tokensNum": 5000
  }'
Example Response: 
{
  "requestId": "81c4198a1d6794503b52134fd77159e2",
  "query": "how to use React hooks for state management",
  "response": "## State Management with useState Hook in React\n\nhttps://www.geeksforgeeks.org/reactjs/state-management-with-usestate-hook-in-react/\n\n```\nimport React, {\n  useState\n} from 'react';\n\nfunction InputField() {\n  const [name, setName] = useState('');\n\n  const handleChange = (event) => {\n    setName(event.target.value);\n  }\n\n  return (\n    <div>\n      Name:\n      <input onChange={handleChange} />\n      Entered name: {name}\n    </div>\n  );\n}\n\nexport default InputField;\n```\n\n## Basic useState Example\n\n```\nimport { useState } from 'react';\n\nfunction Example() {\n  const [count, setCount] = useState(0);\n\n  return (\n    <div>\n      <p>You clicked {count} times</p>\n      <button onClick={() => setCount(count + 1)}>\n        Click me\n      </button>\n    </div>\n  );\n}\n```\n\n## Custom Hook for Counter State Management\n\n```\nimport { useState } from \"react\";\n\nconst useCounter = () => {\n  const [count, setCount] = useState(0);\n\n  const increment = () => {\n    setCount((prevCount) => prevCount + 1);\n  };\n\n  const decrement = () => {\n    setCount((prevCount) => prevCount - 1);\n  };\n\n  return { count, increment, decrement };\n};\n\nexport default useCounter;\n```\n\n...(response continues with more code examples)",
  "resultsCount": 502,
  "costDollars": "{\"total\":1,\"search\":{\"neural\":1}}",
  "searchTime": 3112.290825000033,
  "outputTokens": 4805
}
Library Usage Examples 
curl -X POST 'https://api.exa.ai/context' \
  -H 'x-api-key: YOUR-EXA-API-KEY' \
  -H 'Content-Type: application/json' \
  -d '{
    "query": "pandas dataframe filtering and groupby operations",
    "tokensNum": "dynamic"
  }'
Framework Setup and Configuration 
curl -X POST 'https://api.exa.ai/context' \
  -H 'x-api-key: YOUR-EXA-API-KEY' \
  -H 'Content-Type: application/json' \
  -d '{
    "query": "Next.js 14 app router with TypeScript configuration",
    "tokensNum": "dynamic"
  }'

​
Response Format

The API returns a JSON response with the following structure: 
{
  "requestId": "req_12345",
  "query": "how to use React hooks for state management",
  "response": "// Formatted code snippets and contextual examples\n...",
  "resultsCount": 15,
  "costDollars": "0.0025",
  "searchTime": 1.234,
  "outputTokens": 1247
}

​
Parameters

​
query (required)

  • Type: string
  • Description: Search query to find relevant code snippets
  • Example: "how to use React hooks for state management"
  • Min Length: 1 character
  • Max Length: 2000 characters

​
tokensNum (optional)

  • Type: string | integer
  • Default: "dynamic"
  • Description: Token limit for the response
  • Options:
    • "dynamic": Automatically determine optimal response length
    • 50-100000: Specific number of tokens to return (5000 is good default for most queries, and use 10000 when 5k doesn’t provide enough context)
Token Management
  • Use "dynamic" for most queries to get optimal, token-efficient responses
  • Specify exact token counts when you need precise output length control
  • Higher token counts return more comprehensive examples but cost more

​
Integration Examples

Using with Python 
import requests

def get_code_context(query, tokens="dynamic"):
    response = requests.post(
        "https://api.exa.ai/context",
        headers={
            "Content-Type": "application/json",
            "x-api-key": "YOUR_API_KEY"
        },
        json={
            "query": query,
            "tokensNum": tokens
        }
    )
    
    result = response.json()
    return result["response"]

# Example usage
context = get_code_context("Express.js middleware for authentication")
print(context)
Using with JavaScript/Node.js 
async function getCodeContext(query, tokensNum = "dynamic") {
  const response = await fetch("https://api.exa.ai/context", {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      "x-api-key": "YOUR_API_KEY"
    },
    body: JSON.stringify({
      query,
      tokensNum
    })
  });
  
  const result = await response.json();
  return result.response;
}

// Example usage
const context = await getCodeContext("Svelte component lifecycle methods");
console.log(context);

​
About Exa Code

Vibe coding should never have a bad vibe. exa-code is a huge step towards coding agents that never hallucinate. When your coding agent makes a search query, exa-code searches over billions of GitHub repos, docs pages, Stack Overflow posts, and more, to find the perfect, token-efficient context that the agent needs to code correctly. It’s powered by the Exa search engine.

​
Use with MCP

You can also use exa-code through the Exa MCP server for seamless integration with AI coding assistants like Claude, Cursor, and other MCP-compatible clients. The MCP integration provides the same powerful code context search capabilities directly within your development environment without needing to make direct API calls.
⌘I