SERP API
Get accurate and in real-time results sourced from Google, Bing, and more.
Residential Proxies
Over 100 million real residential IPs from genuine users across 190+ countries.
SCRAPING SOLUTIONS
SERP API
Get accurate and in real-time results sourced from Google, Bing, and more.
Web Scraper API
With 120+ prebuilt and custom scrapers ready for any use case.
Webpage
No blocks, no CAPTCHAs—unlock websites seamlessly at scale.
Scraping Browser
Execute scripts in stealth browsers with full rendering and automation.
PROXY INFRASTRUCTURE
Two Ways In — Choose by Architecture
SDK and MCP share the same Key and data sources — switch or run both as you go from prototype to production.
SDK Pack it into the app
The Thordata SDK wraps into a LlamaIndex FunctionTool, running in the same process as your app.
- Install and go, minimal configuration
- Works directly with LlamaIndex workflows
- Fastest path for prototypes and custom AI apps
- In-process calls, lowest latency
Typical scenarios: developers, RAG applications, AI prototyping
View integration docsMCP Independent Data Provision Service
Search runs on the Thordata MCP Server; LlamaIndex apps connect via adapter, decoupling tools from agents.
- One service, reused across AI apps
- Connects with any MCP-compatible client
- Scales independently, no impact on agent logic
- Supports stdio / HTTP / SSE transports
Typical scenarios: production systems, enterprise apps, multi-agent workflows
View integration docsIntegration Steps
From API Token to your first LlamaIndex app with real-time search — 3 steps.
1
Get your API token
Create a Thordata account for a free token. No credit card required.
2
Choose your integration method
Use the SDK for direct development; choose MCP for cross-app reusable search and service-based deployment.
3
Start building AI applications
Multi-engine structured results in place — retrieval augmentation and agent web access, ready to go.
STEP 1Install SDK (BASH)
pip install llama-index-tools-thordata-serp
STEP 2Call the Search Tool (PYTHON)
import json, os
from llama_index.tools.thordata_serp import ThordataSerpToolSpec
api_key = os.environ["THORDATA_SERP_API_KEY"]
tool_spec = ThordataSerpToolSpec(api_key=api_key)
result = tool_spec.search(query="coffee", params={"num": 10})
print(json.loads(result)["data"]["result"]["organic"][0]["title"])Real Return Example - JSON
1{2"ok": true,3"status": 200,4"engine": "google",5"data": {6 "engine": "google",7 "query": "coffee",8 "taskId": "65f2a1b8c9d0e4f7a1b2",9 "result": {10 "organic": [11 {12 "link": "https://en.wikipedia.org/wiki/Coffee",13 "display_link": "en.wikipedia.org",14 "title": "Coffee",15 "description": "Coffee is a beverage brewed from roasted beans."16 },17 {18 "link": "https://www.britannica.com/topic/coffee",19 "display_link": "britannica.com",20 "title": "Coffee | Origin, Types, Uses, History, & Facts",21 "description": "Coffee, beverage brewed from the roasted seeds of the fruit."22 }23 ],24 "people_also_ask": [25 {26 "question": "How is coffee made?",27 "answers": "Coffee is brewed from roasted coffee beans."28 }29 ]30 }31}32}
Every tool returns a JSON string. The envelope is always ok / status / engine / data; set response_mode="compact" to drop verbose metadata and save context.
Why Choose the Thordata LlamaIndex Integration?
Real-time, reliable, structured search data for your LlamaIndex applications.
One ticket to all engines
One API call to Google, Bing and DuckDuckGo; no per-engine integrations.
Real-time supply for RAG
The latest web information beyond static knowledge bases, fed straight into the LlamaIndex retrieval layer.
Structured out of the box
Clean JSON / Markdown instead of raw pages; zero cleaning in the indexing pipeline.
Zero infrastructure burden
Proxies, crawlers and search infrastructure fully managed; your team just builds the app.
Built for AI Search Workflows
Beyond basic web search — a developer-friendly SERP API for LlamaIndex applications.
- Real search results — Structured delivery of rank, title, snippet and URL, ready for indexing.
- Reliable collection infrastructure — Built-in proxy rotation, CAPTCHA handling and request optimization for stable results.
- Developer-friendly tools — API usage stats and search history available to monitor and optimize AI workflows.
Core Features
Built for LlamaIndex developers — create AI applications with real-time search.
Real-time SERP data
Get results from Google, Bing, DuckDuckGo and other search engines.
Agent & RAG integration
Supports LlamaIndex Tools, SDK, MCP and API access.
Structured output
Provides results in JSON and HTML formats.
Flexible search controls
Supports multi-language search, region, pagination and search type configuration.
Typical Use Cases
See how Thordata helps developers build smarter LlamaIndex applications.
Real-time RAG enhancement
Expand LlamaIndex knowledge systems with live web data; answers go beyond existing documents.
Agents online on demand
FunctionAgent / ReActAgent call search tools autonomously when external information is needed.
Research workflow automation
Drive research workflows with real-time search data; multi-source information auto-summarized.
SEO and content intelligence
Keyword research and content optimization powered by SERP data; topic choices backed by evidence.
Frequently asked questions
How do I integrate Thordata with LlamaIndex?
Two ways: SDK embeds directly into your app process for fast development; MCP runs as an independent service reused across AI apps, for production and multi-agent architectures. Both share the same Key and data sources.
Can Thordata be used in LlamaIndex RAG applications?
Yes. Real-time SERP results serve as a supplementary data source for the RAG retrieval layer, covering the latest web information beyond static documents — structured JSON enters the indexing pipeline with no cleaning.
Why choose Thordata over other search APIs?
Geotargeting across 190+ countries, fully managed proxies and CAPTCHAs, structured JSON / Markdown output, transparent pricing from $0.70 per 1,000 responses, and 500 free responses on sign-up.
Which search engines are supported?
Google, Bing and DuckDuckGo, with geotargeting across 190+ countries and cities, plus multi-language and desktop / mobile parameters.
How do I claim free credits, and what happens when they run out?
500 free responses over 7 days on sign-up, no credit card; after that, pay-as-you-go from $0.70 per 1,000, no prepayment pressure.
Ready to Build Smarter AI Applications?
Use Thordata for free now — add real-time search to your LlamaIndex applications.
No credit card · 7-day refund guarantee