Give your AI agent eyes on the web. Structured search results as JSON. Multi-query support. No search provider API keys needed.
# Search the web $ curl "agent-search.167.148.41.86.nip.io/api/search?q=AI+agent+frameworks+2026" { "query": "AI agent frameworks 2026", "results": [ { "title": "Top AI Agent Frameworks Compared", "url": "https://example.com/agents", "snippet": "Compare CrewAI, LangGraph, AutoGen...", "hostname": "example.com" }, ... ], "count": 10 }
Structured results, multi-query support, and LLM-ready extract mode
GET or POST a query, get structured JSON results with title, URL, snippet, and hostname. Sub-second response times.
Search up to 5 queries in a single request. Perfect for research pipelines, comparison tasks, and parallel information gathering.
Get search results formatted as numbered text — ideal for feeding directly to LLMs as context without extra parsing.
No Google API key, no Bing API key, no SerpAPI subscription. Just call the endpoint. We handle everything.
JSON responses, Bearer token auth, machine-discoverable via /.well-known/agent.json and /llms.txt. Built for automated use.
30 searches per day free, no API key required. Scale with USDC on Base when you need more. No credit card ever.
Try it now — real search, real results
Simple, predictable REST API
Drop-in code for your language
# Simple search curl "http://agent-search.167.148.41.86.nip.io/api/search?q=best+crypto+wallets&max=5" # Multi-query search (research multiple topics at once) curl -X POST http://agent-search.167.148.41.86.nip.io/api/search/multi \ -H "Content-Type: application/json" \ -d '{"queries":["langchain agents","autogen framework","crewai vs langgraph"]}' # Extract mode (LLM-ready text output) curl -X POST http://agent-search.167.148.41.86.nip.io/api/search/extract \ -H "Content-Type: application/json" \ -d '{"query":"how to build AI agents","maxResults":5}'
import requests API = "http://agent-search.167.148.41.86.nip.io" # Simple search r = requests.get(f"{API}/api/search", params={"q": "best crypto wallets 2026", "max": 5}) for result in r.json()["results"]: print(f'{result["title"]}: {result["url"]}') # Multi-query for research pipelines r = requests.post(f"{API}/api/search/multi", json={ "queries": ["langchain agents", "crewai", "autogen"] }) for qr in r.json()["results"]: print(f"Query: {qr['query']}, Found: {qr['count']}") # Extract mode — feed directly to LLM r = requests.post(f"{API}/api/search/extract", json={ "query": "AI agent security best practices" }) context = r.json()["extracted"] # context is numbered text ready for LLM input
const API = "http://agent-search.167.148.41.86.nip.io"; // Simple search const { results } = await fetch( `${API}/api/search?q=best+AI+tools&max=5` ).then(r => r.json()); results.forEach(r => console.log(r.title, r.url)); // Multi-query for parallel research const multi = await fetch(`${API}/api/search/multi`, { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify({ queries: ["langchain", "crewai", "autogen"] }) }).then(r => r.json()); // Extract mode for LLM context const { extracted } = await fetch(`${API}/api/search/extract`, { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify({ query: "AI agent security" }) }).then(r => r.json()); // Feed extracted text to your LLM const messages = [ { role: "user", content: `Based on:\n${extracted}\n\nSummarize.` } ];
Start free. No credit card required. Scale with USDC.
35+ APIs built for AI agents. All with free tiers. All accepting USDC.