Remote MCP

Valyu MCP Server enables AI assistants like Claude to perform real-time searches through the Valyu Deepsearch API, allowing them to access up-to-date information from the web, academic journals and textbooks, and financial data. It is open-source, checkout GitHub.

​

Remote Valyu MCP

Connect directly to Valyu’s hosted MCP server (instead of running it locally).

​

Remote Valyu MCP URL

https://mcp.valyu.network/mcp?valyuApiKey=your-valyu-api-key

Get your API key from platform.valyu.network/user/account/apikeys.

​

Configuration

​

​

Claude Desktop Configuration

Add this to your Claude Desktop configuration file:

{
  "mcpServers": {
    "valyu": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.valyu.network/mcp?valyuApiKey=your-valyu-api-key"
      ]
    }
  }
}

​

Claude Code Configuration

Add the Valyu MCP server to Claude Code using the CLI:

claude mcp add --transport http valyuMcp "https://mcp.valyu.network/mcp?valyuApiKey=YOUR_API_KEY"

​

Install on Cursor

​

OpenAI Responses API

This is how to try out the Valyu MCP server in the OpenAI playground.

Easily hook up the Valyu MCP server to the OpenAI Responses API:

from openai import OpenAI
client = OpenAI()

response = client.responses.create(
  model="gpt-4.1",
  input=[],
  text={
    "format": {
      "type": "text"
    }
  },
  reasoning={},
  tools=[
    {
      "type": "mcp",
      "server_label": "Valyu",
      "server_url": "https://mcp.valyu.network/mcp?valyuApiKey=VALYU_API_KEY",
      "allowed_tools": [
        "valyu_search",
        "valyu_web_search",
        "valyu_academic_search",
        "valyu_financial_search"
      ],
      "require_approval": "always"
    }
  ],
  temperature=1,
  max_output_tokens=2048,
  top_p=1,
  store=True
)

​

Available Tools

Valyu MCP includes four specialized search tools optimized for different types of information:

​

Tool Parameters

​

valyu_search

• query (required): Search query string
• search_type: 'all' (default), 'proprietary', or 'web'
• max_num_results: 1-50 results (default: 10)
• max_price: Maximum cost per 1000 retrievals in dollars (default: 20)
• relevance_threshold: 0.0-1.0 relevance score threshold (default: 0.5)
• included_sources: Array of specific data sources
• included_domains: Comma-separated domains for web search
• start_date: Filter from date (YYYY-MM-DD)
• end_date: Filter until date (YYYY-MM-DD)
• category: Content category filter

​

valyu_web_search

• query (required): Web search query
• max_num_results: 1-50 results (default: 10)
• included_domains: Comma-separated domains (e.g., ‘cnn.com,bbc.com’)
• start_date: Filter from date (YYYY-MM-DD)
• end_date: Filter until date (YYYY-MM-DD)

​

valyu_academic_search

• query (required): Academic research query
• max_num_results: 1-20 results (default: 10)
• included_sources: Academic sources (e.g., [‘valyu/valyu-arxiv’, ‘valyu/valyu-pubmed’])
• start_date: Filter from date (YYYY-MM-DD)
• end_date: Filter until date (YYYY-MM-DD)
• category: Research category (e.g., ‘machine learning’, ‘finance’)

​

valyu_financial_search

• query (required): Financial query (e.g., ‘AAPL stock price’, ‘Bitcoin volatility’)
• max_num_results: 1-50 results (default: 10)
• included_sources: Financial sources (e.g., [‘valyu/valyu-stocks-US’, ‘valyu/valyu-crypto’])
• start_date: Filter from date (YYYY-MM-DD)
• end_date: Filter until date (YYYY-MM-DD)

​

Response Format

All tools return structured responses with essential fields:

• title: Content title or headline
• url: Source URL
• content: Main content (text or structured data)

For financial queries, responses include structured data with OHLC values, timestamps, and market metrics.

​

Usage Examples

Once configured, you can ask your AI assistant to perform various types of searches:

​

General Research

“Search for recent developments in renewable energy storage technologies”

​

Academic Research

“Find academic papers published in 2024 about machine learning applications in drug discovery”

​

Web Search

“Search for recent news about artificial intelligence regulations from government websites”

​

Financial Analysis

“Get Tesla stock performance data for the last quarter and analyze the trends”

​

Domain-Specific Searches

“Search Bloomberg and Reuters for recent analysis on cryptocurrency market trends”

​

Date-Filtered Research

“Find academic papers from 2023-2024 about climate change mitigation strategies”

​

Features

• Multi-source Integration: Access academic papers, real-time web content, and financial data from a single interface
• Advanced Filtering: Date ranges, domain restrictions, source selection, and relevance thresholds
• Cost Management: Configurable price limits to control API usage costs
• High Relevance: Tunable relevance thresholds ensure quality results
• Structured Financial Data: Time-series data with OHLC values for market analysis
• Real-time Information: Up-to-date web content and current market data
• OpenAI Compatible: Direct integration with OpenAI’s Responses API

​

Transport Protocol

Uses HTTP/SSE transport (JSON-RPC over HTTP with Server-Sent Events) for:

• ✅ Direct OpenAI Responses API compatibility
• ✅ Real-time streaming responses
• ✅ Stateless authentication via URL parameters
• ✅ Standard MCP protocol compliance

​

Pricing

Valyu MCP operates on a usage-based pricing model through Valyu’s Deepsearch API. Configure max_price parameters to control costs per search operation. Visit Valyu Pricing for current rates.

​

Additional Resources

​

Support

• Documentation: docs.valyu.network
• API Keys: platform.valyu.network/user/account/apikeys
• Platform: platform.valyu.network