OroIntelDocs
OpenAPISign up now
For agents:start at/llms.txt·everything as Markdown at/llms-full.txt·or add theMCP server.

Use Oro Intel with LangChain

Give a LangChain (LangGraph) agent live UK tender and contract data by connecting Oro Intel's remote MCP server as tools — no local server to run.

Prerequisites

  • An Oro Intel API key in ORO_API_KEY (create one, 250 free credits).
  • An LLM key (this example uses Anthropic via ANTHROPIC_API_KEY).
  • pip install langchain-mcp-adapters langgraph "langchain[anthropic]"

Code

agent.py
"""LangChain agent with live UK procurement data via the Oro Intel remote MCP server."""

import asyncio
import os

from langchain_mcp_adapters.client import MultiServerMCPClient
from langgraph.prebuilt import create_react_agent

ORO_API_KEY = os.environ["ORO_API_KEY"]  # https://app.oro-intel.com/dashboard/developers

client = MultiServerMCPClient(
    {
        "oro-intel": {
            "transport": "streamable_http",
            "url": "https://api.oro-intel.com/mcp/",  # trailing slash required
            "headers": {"Authorization": f"Bearer {ORO_API_KEY}"},
        }
    }
)


async def main() -> None:
    tools = await client.get_tools()
    agent = create_react_agent("anthropic:claude-sonnet-5", tools)
    result = await agent.ainvoke(
        {
            "messages": [
                (
                    "user",
                    "Look up the UK company 'Serco' and give me its full supplier "
                    "profile. Report how many public contracts it has won, and the "
                    "credits_charged and credits_remaining from the tool responses.",
                )
            ]
        }
    )
    print(result["messages"][-1].content)


asyncio.run(main())

Expected output

output (trimmed)
Serco Limited (company number 02048608) has won 132 public contracts.
The lookup cost 5 credits and the full profile 12 credits
(credits_charged: 12, credits_remaining: 233).

The agent calls oro_lookup_company (5 credits) then oro_company_profile (12 credits). Every tool response includes credits_charged/credits_remaining, and each tool's description states its cost, so the agent knows before it calls.

Next steps