MCP Server by handelsregister.ai: Seamless AI Integration with Commercial Register Data

New feature: MCP Server (beta) – seamless AI integration now available!

handelsregister.ai is expanding its platform with an exciting new feature: the MCP Server (Model Context Protocol) in open beta. It makes integrating Handelsregister (commercial register) data into your AI-powered workflows and agent-based systems easier than ever.

What does that mean in practice? Instead of laboriously retrieving data through a REST API and feeding it into your AI applications manually, modern AI models can now access our Handelsregister data on their own: standardized, secure and in real time. The result: faster development, less manual work and more relevant answers from your AI systems, because they can draw directly on current company data.

Below you'll learn what the Model Context Protocol (MCP) is all about, what benefits the new MCP server offers, how the integration works and which use cases it is particularly well suited for.

What is the Model Context Protocol (MCP)?

The Model Context Protocol (MCP) is an open standard introduced in 2024, designed to connect AI models seamlessly with external data sources, tools and services. You can think of MCP as a kind of "USB-C for AI applications". Just as USB-C provides a uniform port for connecting different devices and peripherals, MCP provides a uniform way for AI systems (e.g., ChatGPT or Claude) to access a wide variety of external systems.

Traditionally, AI models were largely isolated from current data: anything that was not in their training set remained hidden from them. Every new data source required a purpose-built "connector" or special implementations. MCP removes this bottleneck by creating a universal protocol: you can make data sources accessible through so-called MCP servers, while AI applications act as MCP clients and connect to these servers. This eliminates the tedious work of adapting to countless different APIs; modern AI models "speak" MCP out of the box and use it to access approved data directly.

Why MCP? Advantages over the REST API

You may be wondering how this differs from using our REST API. In short: MCP is made specifically for AI models, whereas REST was designed more for you as a developer.

• Self-describing interface: An MCP server automatically describes its available functions ("tools") and data structures. The AI model knows in advance which queries it can make and in what format the answers will come. With a REST API, you would have to build this logic into your code manually.
• Standardized calls: Even though MCP servers often still use REST or gRPC in the background, MCP wraps everything in a uniform format. To the model, all tools look the same, whether it's a database, a web service or a file system. You no longer have to deal with different endpoints, parameters or authentication methods.
• Direct use by the AI model: With MCP, the model can decide for itself when to start a Handelsregister lookup. You don't have to orchestrate every single request; instead, the model uses the server as a tool within its reasoning process.
• Fewer hallucinations, more facts: Because the model queries real data when needed, the risk of wrong answers drops. Instead of using outdated training knowledge, it can, for example, fetch current financial figures live from the Handelsregister.
• Simpler workflows: In multi-agent systems, you would otherwise build separate integrations for every step. With MCP, your agent simply sees various tools and decides for itself which ones to use, which saves you a lot of development time.

In short: MCP complements the classic REST API but does not replace it. Existing code can keep using REST, but for AI-based applications MCP opens up a much more direct and elegant path to the data.

Main features of the MCP server

The new MCP Server by handelsregister.ai offers you:

• Universal compatibility: Works with all MCP-compatible clients, whether OpenAI or Anthropic Claude.
• Real-time data access: Access current German company data directly, at any time.
• OpenAI SDK support: Use the server directly in the OpenAI Python SDK.
• Workflow automation: Ideal for multi-agent systems in which Handelsregister data is analyzed and processed automatically.

How to connect the MCP server

Getting started is easy: all you need is your handelsregister.ai API key and an MCP-capable AI client. Here is an example with the OpenAI Python SDK:

    import os
    from openai import OpenAI

    client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

    response = client.responses.create(
        model="gpt-5",
        input="Give me the latest financial data for Konux GmbH from Munich",
        tools=[
            {
                "type": "mcp",
                "server_label": "handelsregister",
                "server_url": "https://mcp.handelsregister.ai/mcp",
                "require_approval": "never",
                "headers": {
                    "X-API-Key": os.environ["HANDELSREGISTER_API_KEY"]
                },
            }
        ],
    )

    print(response.output_text)

In this example, you ask the model a question and it uses the MCP server as a tool on its own. You don't have to program a separate REST call; the model takes care of it.

Authentication and endpoint

The server URL is: https://mcp.handelsregister.ai/mcp
Authentication works as usual via your API key in the header: X-API-Key.

Requirements

• An MCP-compatible client (e.g., OpenAI SDK or Claude Desktop)
• A valid API key from handelsregister.ai
• Python 3.8+ (if you use the SDK example)

Use cases

With the MCP server, you can implement many exciting scenarios:

• Company analyses: Have an AI agent compile reports on annual financial statements, shareholder lists or management changes live.
• Due diligence: Automate the review of company data for investments or acquisitions.
• Multi-agent workflows: Build agent systems that monitor Handelsregister entries and analyze anomalies immediately.
• Integration into pipelines: Add current Handelsregister data to existing dashboards, chatbots or CRM systems.

Beta phase, costs and outlook

The MCP Server is currently in open beta. This means we are actively developing it and look forward to your feedback.

Costs: Usage is billed at the same credit prices as the REST API. There are no additional fees. The rate limits are identical too.

Security: Your API key naturally stays secret and is only transmitted encrypted.

Conclusion

With the MCP Server (beta), you can connect your AI applications directly to Handelsregister data, with no detours and no extra effort. You save development time, and your AI delivers better-founded answers because it accesses current company data.

Try the MCP server and see for yourself what new possibilities it opens up for you. We look forward to your ideas and feedback!

Further guides

• Company data via API: instant access to the German commercial register with handelsregister.ai
• Viewing a GmbH's annual financial statements: 3 ways – free, structured and with export to Excel