Same logs. Familiar tools. A new way to dig into what’s going on.
If you’ve been waiting for something new from Loggly, this one’s for you. We’ve got a new feature designed to help you make sense of your logs using the AI tools you already leverage.
Meet the Loggly MCP Server. It connects your Loggly data to Claude, GitHub Copilot and Cursor so you can ask questions about your logs and get responses right in your AI assistant.
Why did we build this? Because your troubleshooting to-do list probably doesn’t need “learn another tool” on it. You’ve already got the logs in Loggly. Now, bring them into the AI tools that are already part of your kit. No moving your logs to a new platform. No separate AI interface to get used to. Just a new connection that helps you do more with the Loggly you know and love.
MCP? What Does That Mean?
MCP stands for Model Context Protocol. The Loggly MCP Server gives your AI assistant a way to query your Loggly data and bring the results back into the conversation.
Query syntax has its place. But sometimes you just want to ask:
What errors have occurred in the past hour?
Or get a little more specific:
Show me the last 20 errors from the payments service.
Or follow the breadcrumbs:
Find the log events for correlation ID abc-123.
Your connected assistant uses the server to search Loggly and return results. You bring the question. Your logs bring the context.
Put Those Logs to Work
Here are a few places to start:
- Spot the repeat offenders: Explore recent errors grouped by frequency and see what keeps showing up.
- Follow a hunch: Retrieve related log events using a correlation ID.
- Dig into the details: Discover parsed fields, compare field-value counts, and get statistics for numeric fields.
Get useful context from the data you already collect, through tools you already use.
A Little Setup. Then Get Digging.
The Loggly MCP Server is self-hostable. Run it locally or in the infrastructure you manage, connect it using your Loggly reading API token and subdomain, and configure your compatible AI client.
Once connected, you can start asking questions through your assistant. The server’s tools are read-only: they search and analyze logs, rather than change your Loggly data. Limits on query time ranges and result sizes help keep searches bounded.
One important housekeeping item: Always store API tokens securely, not in chat prompts. Follow your organization’s policies for sharing log data with AI tools, too.
Got a Question for Your Logs? Try It Out.
Pick a question you’d normally search manually for in your logs. Ask it in your connected AI assistant. See what your logs have to say.
Want to give it a try? Visit our documentation site for setup instructions.
Then drop into the comments and tell us what you tried. Which assistant did you connect? What did you ask? We’d love to hear how you’re putting this new connection to work.