Announcing the Vantage MCP Server: Use AI to Analyze Your Cost and Usage Data

Today, Vantage is announcing the launch of its open-source, local Model Context Protocol (MCP) Server, which allows customers to seamlessly interact with Vantage through Large Language Model (LLM) tools like Claude, Cursor, and soon ChatGPT, as well as open source agents like Goose and 5ire that act as MCP Clients. Customers can use the MCP Server to incorporate GenAI into their FinOps workflows. For example:

• Ask questions about cloud cost and usage in natural language.
• Perform complex, multi-step analysis and generate actionable recommendations
• Combine the Vantage MCP Server with other MCP’s to create intelligent workflows.

MCP servers are lightweight, standardized interfaces that enable AI assistants to interact with external data sources and tools. By acting as a bridge between AI models and data sources that were historically isolated from LLMs, MCP servers allow for real-time, context-rich interactions without the need for custom integrations. This standardized approach allows organizations to connect AI systems with data sources with a single protocol without the need to create custom code or logic.

Now, with the launch of the Vantage MCP Server, customers can expose their cloud cost and usage data to AI agents in order to automate workflows and perform intelligent analysis. The server acts as a wrapper around Vantage’s existing APIs, allowing users to ask questions about their costs, such as “What was our AWS spend last month?”, or have AI agents perform more complex analysis such as “What duplicate tags do I have that I should standardize as Virtual Tags?”, and receive immediate responses. The Vantage MCP Server can also interact with other MCP servers to expedite analysis of your cost and usage, such as using with the Github MCP Server to tie Cost Anomalies in Vantage to Github pull requests.

The Vantage MCP Server version 0.0.1 is now available for all Vantage customers. To get started, visit the Vantage MCP Server GitHub repository for setup instructions and integration guidelines. FinOps practitioners attending FinOps X can also learn more about MCP use in FinOps at the session Connecting AI to Your FinOps Data: APIs, MCPs, and Real-world Insights from Block with Youssef Ibrahim (Head of Infrastructure Cost Efficiency at Block) and Brooke McKim (Co-Founder & CTO at Vantage).

Frequently Asked Questions

1. What is being launched today?

Vantage is launching an open-source MCP Server, which allows customers to use AI assistants to access and interact with their organization’s data within Vantage, enabling more informed and context-aware AI responses.

2. Who is the customer?

This feature is available to all Vantage users, specifically those interested in using GenAI tools in their FinOps workflows.

3. How much does this cost?

There is no additional cost to for customers to use the Vantage MCP Server.

4. Why are MCP servers useful? MCP servers provide a standard way for AI assistants to access and interact with external data sources. Teams can use MCP servers to interact with internal tools and datasets, such as Vantage cost data, using natural language.

The Vantage MCP Server bridges AI tools with Vantage’s cloud cost management capabilities

5. What functionality does v0.0.1 of the Vantage MCP Server have?

The Vantage MCP Server v0.0.1 is currently read-only, meaning it cannot create new resources. The server has the ability to:

• Get cost data, either from a specific cost report or any customizable period via the /v2/costs API endpoint.
• List out existing cost reports
• List all cost provider integrations (e.g., AWS, Azure, GCP) available to provide costs data from and their associated accounts
• List tag keys and tag values (both from integrated providers and Vantage Virtual Tags)
• List cost anomalies that were detected on cost reports.

Additional functionality will continually be added to the Vantage MCP Server. See the GitHub ReadMe file for most updated functionality.

6. What are examples of how I can use the Vantage MCP Server to analyze my cloud costs?

Some examples of ways you can leverage the Vantage MCP server include

• Ask natural language questions like “How much did my team spend on EC2 last month?” or “What percent did my S3 costs increase?” and get instant answers from your cloud cost data.
• Use the Vantage MCP Server to automatically surface missing or messy tags and unallocated resources, eliminating the tedious work of identifying them yourself.
• Interact with multiple MCPs at once to compare insights, share context, and troubleshoot faster, no need to jump between tools or sessions.

7. How does the Vantage MCP Server integrate with my Vantage environment?

The server integrates with your environment using a read-only Vantage API token. To prevent possibility of writes to your account, API tokens with write access will currently throw an error. It is recommended that you create a dedicated API token to manage your authentication with the Vantage MCP server. Follow the API documentation for instructions on how to create a Vantage API token or directly create one from account settings.

8. Does Vantage have access to see what prompts I ask an LLM?

Vantage does not receive any information about the prompt you ask an LLM agent.

9. How do I get started running the Vantage MCP Server locally?

Follow the instructions provided in the GitHub repository’s README. We recommend creating a separate read-only API token for use with the MCP server.

10. What are the supported MCP Clients I can use this with?

The Vantage MCP Server is compatible with any AI Agents that provide MCP Client functionality. At the time of this blog post, examples of supported clients include Anthropic’s Claude Desktop application, Cursor IDE, and Block’s local AI agent Goose. ChatGPT supports MCP through their SDK, however does not have desktop application support yet. See documentation from each provider to understand minimum version requirements required to run MCP hosts for each. You can find a list of example clients in the MCP documentation.

11. Is there a hosted Vantage MCP Server (SSE Transport) available?

At this time, only a locally run (Standard Input/Ouput [stdio] Transport) MCP server is available.

12. Can I contribute to this project?

Yes, Vantage is taking contributions to this project through Pull Requests in the GitHub repository. See the contribution guide for more information.

13. If I need support using the Vantage MCP Server, how can I get that?

Through our regular support channels - please email support@vantage.sh and/or reach out to your Customer Success Manager where appropriate.

14. I would like to provide feedback or talk more on this. How can I engage more with the Vantage team on MCP related work?

In the Vantage Community Slack, there is a new #mcp channel where you can provide feedback and interact with Vantage employees contributing to the project, as well as interact with other Vantage users on how they are utilizing the Vantage MCP Server. The Vantage Community Slack is a free community for anyone to join.

15. What license is the Vantage MCP Server released under, and how can I contribute?

For commercial purposes, you need a private agreement with Vantage. Please contact partnerships@vantage.sh. For non-commercial purposes, the Vantage MCP Server is open source and available under the MIT License. Contributions are welcome; you can fork the repository, open a pull request, and help improve the project. Please submit any bug reports or issues through the GitHub Issues page.