Searlo

SERP API Built for AI Agents & LLMs

Ground your AI with real-time web data. Optimized for LLMs with TOON format, sub-second latency, and streaming support.

Start BuildingView AI Docs
GET/api/v1/search/web?q=postgres+connection+pooling&num=3
200 OK0.17s1 credit85,000 results
{  "searchTime": 0.17,  "totalResults": 85000,  "page": 1,  "organic": [    {      "position": 1,      "title": "31.10. Connection Pools and Data Sources",      "link": "www.postgresql.org/docs/…",      "domain": "postgresql.org",      "snippet": "For an environment without an application se…"    },    { "position": 2, "domain": "medium.com", … },    { "position": 3, "domain": "stackoverflow.blog", … }  ],  "credits": 1}

60% Fewer Tokens with TOON Format

Traditional JSON wastes tokens on quotes, brackets, and verbose keys. TOON delivers the same information in a fraction of the tokens.

JSON: 847 tokens

TOON: 312 tokens

query: best laptops 2025 intent: commercial results[3]{title,url,score}: MacBook Pro Review,apple.com,0.98 Dell XPS Guide,dell.com,0.95 ThinkPad X1,lenovo.com,0.91

60% Fewer Tokens with TOON Format

query: best laptops 2025
intent: commercial
results[3]{title,url,score}:
 MacBook Pro Review,apple.com,0.98
 Dell XPS Guide,dell.com,0.95
 ThinkPad X1,lenovo.com,0.91

Built for AI Developers

  • TOON Format

    Token-Optimized Object Notation reduces token usage by 60%+ compared to JSON, saving costs on GPT-4, Claude, and Gemini calls.

  • Sub-second Latency

    Real-time responses that won't slow down your AI agent. 10x faster than SerpApi for conversational interfaces.

  • MCP Protocol Support

    Official Model Context Protocol integration. Connect Searlo to Claude, Cursor, Windsurf, or any MCP-compatible AI assistant. See our dedicated MCP Server page for setup guides.

  • Streaming + Intent Detection

    Stream results as they arrive. Built-in query intent classification helps your AI understand user needs better.

Works With Your Stack

Easy integration with popular AI frameworks and tools.

LangChain Use as a tool in your LangChain agents from langchain.tools import SearloSearch tool = SearloSearch(api_key="...") agent = Agent(tools=[tool])

Features

  • LangChain

    Use as a tool in your LangChain agents

Works With Your Stack

from langchain.tools import SearloSearch

tool = SearloSearch(api_key="...")
agent = Agent(tools=[tool])

OpenAI Functions Perfect for function calling with GPT-4 functions = [{ "name": "web_search", "description": "Search the web", "parameters": {...} }]

Features

  • OpenAI Functions

    Perfect for function calling with GPT-4

Get started in one request

functions = [{
  "name": "web_search",
  "description": "Search the web",
  "parameters": {...}
}]

RAG Pipeline Augment retrieval with fresh web data # Combine with your vector DB web_results = searlo.search(query) context = vectordb.query(query) response = llm.generate(context + web_results)

Features

  • RAG Pipeline

    Augment retrieval with fresh web data

Get started in one request

# Combine with your vector DB
web_results = searlo.search(query)
context = vectordb.query(query)
response = llm.generate(context + web_results)

Use Cases

  • AI Assistants & Chatbots

    Give your AI assistant access to real-time information. Answer questions about current events, prices, and more.

  • RAG Applications

    Augment your retrieval pipeline with fresh web data. Combine vector search with real-time SERP results.

  • Research Agents

    Build autonomous research agents that can explore topics across the web and synthesize findings.

  • Content Generation

    Generate content grounded in real facts. Verify claims, find sources, and add citations.

Features

  • What is TOON format?

    TOON (Token-Optimized Object Notation) is our proprietary format designed for LLMs. It reduces token count by 60%+ compared to JSON while preserving all information. This means lower costs and faster processing for your AI applications.

  • How do I integrate with LangChain?

    We provide a LangChain-compatible tool that you can add to your agents. Simply pip install our SDK and use the SearloSearch tool. Full documentation and examples are available in our docs.

  • Is streaming supported?

    Yes! Use the stream=true parameter to receive results as they're processed. This is ideal for real-time chat interfaces where you want to show progress.

  • How fresh is the data?

    Our results are real-time - you get what users would see if they searched Google right now. No caching, no stale data. Perfect for AI that needs current information.

Ready to supercharge your AI?

Start free with 3,000 free credits. No credit card required.

Related

From the team behind Searlo

We also build RenderKit — the browser-rendering API for screenshots, PDFs, and LLM-ready markdown. Same billing model, same developer-first API design.

Explore the full RenderKit API →

Features

FAQ

FAQ

What is TOON format?

TOON (Token-Optimized Object Notation) is our proprietary format designed for LLMs. It reduces token count by 60%+ compared to JSON while preserving all information. This means lower costs and faster processing for your AI applications.

How do I integrate with LangChain?

We provide a LangChain-compatible tool that you can add to your agents. Simply pip install our SDK and use the SearloSearch tool. Full documentation and examples are available in our docs.

Is streaming supported?

Yes! Use the stream=true parameter to receive results as they're processed. This is ideal for real-time chat interfaces where you want to show progress.

How fresh is the data?

Our results are real-time - you get what users would see if they searched Google right now. No caching, no stale data. Perfect for AI that needs current information.

From the team behind Searlo

We also build RenderKit — the browser-rendering API for screenshots, PDFs, and LLM-ready markdown. Same billing model, same developer-first API design.

URL to Markdown APIClean markdown plus token-aware chunks for RAG ingestion.MCP web scraping serverGive Claude or Cursor a page-to-markdown tool over MCP.MCP screenshot serverLet an agent see the page it is reasoning about.Token-aware chunkingWhy naive character splits wreck retrieval quality.

Explore the full RenderKit API →

Related