Free Web Search API — Search the Web via REST

{
  "query": "best javascript frameworks 2026",
  "engine": "duckduckgo-html",
  "results": [
    {
      "title": "Top 10 JavaScript Frameworks in 2026 - Dev.to",
      "url": "https://dev.to/top-javascript-frameworks-2026",
      "snippet": "A comprehensive comparison of React, Vue, Svelte, Solid, and more...",
      "hostname": "dev.to"
    },
    {
      "title": "JavaScript Framework Benchmarks 2026",
      "url": "https://krausest.github.io/js-framework-benchmark/",
      "snippet": "Performance comparison across 50+ frameworks with real DOM...",
      "hostname": "krausest.github.io"
    }
  ],
  "count": 10,
  "noResults": false
}

curl -X POST https://agent-gateway-kappa.vercel.app/v1/agent-search/api/search/extract \
  -H "Content-Type: application/json" \
  -d '{"query": "latest AI regulations Europe 2026", "maxResults": 5}'

{
  "query": "latest AI regulations Europe 2026",
  "engine": "duckduckgo-html",
  "results": [...],
  "count": 5,
  "text": "[1] EU AI Act Implementation Timeline\nhttps://...\nThe European Union's AI Act enters full enforcement...\n\n[2] AI Governance in 2026\nhttps://...\nNew guidelines for high-risk AI systems...\n\n..."
}

import requests

SEARCH_API = "https://agent-gateway-kappa.vercel.app/v1/agent-search/api/search/extract"
LLM_API = "https://agent-gateway-kappa.vercel.app/v1/agent-llm/api/chat"

def search_and_answer(question, provider_key):
    """Search the web, then answer using an LLM with search context."""

    # Step 1: Search the web
    search = requests.post(SEARCH_API, json={
        "query": question,
        "maxResults": 5
    }).json()

    context = search.get("text", "No results found.")

    # Step 2: Ask LLM with search context
    answer = requests.post(LLM_API, json={
        "model": "gpt-4o-mini",
        "messages": [
            {"role": "system", "content": f"Answer based on these search results:\n\n{context}"},
            {"role": "user", "content": question}
        ],
        "provider_key": provider_key,
        "max_tokens": 500
    }).json()

    return {
        "question": question,
        "answer": answer["choices"][0]["message"]["content"],
        "sources": [r["url"] for r in search.get("results", [])],
        "tokens": answer.get("usage", {}).get("total_tokens", 0)
    }

# Usage
result = search_and_answer(
    "What are the best AI coding assistants in 2026?",
    provider_key="sk-your-openai-key"
)

print(f"Q: {result['question']}")
print(f"A: {result['answer']}")
print(f"\nSources:")
for url in result["sources"]:
    print(f"  - {url}")

import requests
import json

SEARCH_URL = "https://agent-gateway-kappa.vercel.app/v1/agent-search/api/search"

def web_search(query: str, max_results: int = 5) -> str:
    """Search the web and return formatted results.

    Use this tool when you need current information from the web,
    to verify facts, or to find URLs for further research.
    """
    resp = requests.get(SEARCH_URL, params={"q": query, "max": max_results})
    data = resp.json()

    if data.get("noResults"):
        return f"No results found for: {query}"

    results = []
    for r in data["results"]:
        results.append(f"- {r['title']} ({r['hostname']})\n  {r['url']}\n  {r['snippet']}")

    return f"Search results for '{query}':\n\n" + "\n\n".join(results)

# Use in any agent framework:
# tool = Tool(name="web_search", func=web_search, description="Search the web")
#
# Or call directly:
print(web_search("latest Python 3.14 features"))