Author
Mano Toth
Senior Technical Writer
We’ve expanded Axiom MCP Server to work seamlessly as a remote connection, making it easier than ever to bring your data into AI agents like Claude and Cursor. Additionally, Axiom MCP Server now supports a set of powerful prompts designed to accelerate common observability workflows and take you from raw events to insights faster.
Remote MCP Server access
Until now, using MCP with Axiom meant running a local server. That worked well for developers, but it added setup steps and maintenance overhead. The new remote MCP option lets you connect in seconds, so you can start asking your AI assistant meaningful questions about your data right away.
How remote MCP Server works:
- Remote server access at
https://mcp.axiom.co/mcpwith zero local dependencies. - Seamless authentication via your browser, with the option to revoke access anytime.
- Works with any MCP-compatible agent, so you can query and explore your Axiom data from your preferred AI assistant.
Whether you’re troubleshooting an incident, checking monitor history, or exploring new datasets, connecting remotely lets you focus on the questions, not the setup.
Supported MCP prompts
In addition to core MCP tools like schema exploration, saved query retrieval, and APL execution, the server now supports the following domain-specific prompts:
- Correlate events across datasets
- Investigate data quality issues
- Detect anomalies in events
- Establish performance baselines
- Explore unknown datasets
- Analyze monitor health
Together, these prompts help you move from raw data to actionable insights faster, whether you’re debugging production issues, analyzing anomalies, or validating the health of your monitors.
Get started
Set up a remote connection in your AI agent and authenticate once in your browser to start exploring. For details, see the Axiom MCP Server documentation.