Claude + New Relic: What the Integration Can (and Can't) Do in 2026
Yes — New Relic ships an official remote MCP server for Claude. New Relic hosts its own MCP server at https://mcp.newrelic.com/mcp/ (US) and https://mcp.eu.newrelic.com/mcp/ (EU), and you connect it to Claude as a custom connector — it isn’t a one-click entry in Anthropic’s connectors directory, but the server itself is built and hosted by New Relic. Once it’s wired up, Claude can run NRQL queries, inspect entities and dashboards, and look at alerts and incidents inside a chat. The catch is the one every Claude MCP setup carries: it only works inside a conversation you start. There are no triggers, nothing watches your telemetry for you, and nothing fires while you’re away.
Here’s exactly what the integration does, how to turn it on, where the limits bite, and what to use if you want New Relic work that runs on its own.
What the New Relic MCP server does
New Relic announced its agentic MCP server on Nov 4, 2025 and it became broadly available through early 2026 (New Relic still labels parts of it pre-release, so expect the odd rough edge). It connects Claude to your New Relic account and exposes your observability data as tools you drive with natural language.
In practice, the New Relic MCP server lets Claude:
- Run NRQL queries — ask questions across your metrics, events, logs, and traces and get answers back in the chat.
- Inspect entities and dashboards — pull up services, hosts, and the dashboards you’ve built.
- Look at alerts and incidents — review open incidents, alert conditions, and what fired.
- Investigate with least privilege — the docs recommend scoped RBAC accounts, and tool access is filtered with
include-tagsheaders (discovery, alerting, data-access), so you decide how much the connection can touch.
The everyday wins: “why did latency spike on the checkout service in the last hour,” “show me error rates by host today,” “what alerts fired overnight.” All without leaving Claude or hand-writing NRQL.
Don’t confuse the two directions here: New Relic can also monitor AI agents you build (its Agentic AI Monitoring product), which is the opposite flow — Claude querying New Relic is the MCP server described above.
How to set it up
Setup is a remote custom connector:
- Grab a New Relic API key (or use the recommended OAuth flow) from your New Relic account — a scoped, least-privilege user is the sane default.
- In Claude, go to Settings → Connectors → Add custom connector and point it at
https://mcp.newrelic.com/mcp/(or the EU endpoint). - Authorize with OAuth or your API key. Optionally set
include-tagsheaders to limit which tool groups load. - Back in a chat, ask Claude to query your telemetry — it’ll route through the MCP server to NerdGraph.
New Relic’s docs also cover setup for Claude Code and Claude Desktop, alongside other MCP clients. Adding a remote custom connector on claude.ai requires a paid Claude plan; the local setup runs through Claude Desktop or Claude Code.
The limits that actually matter
The MCP server is good at pulling New Relic into a conversation. But its shape is “an SRE you query,” not “an agent that watches.” Three limits define it:
- No triggers, no monitoring. The server only works inside a conversation you start. There’s no “when error rate crosses a threshold, page the on-call and open an incident” or “summarize every alert into Slack as it fires.” Nothing runs on an event — you have to be there, prompting.
- Conversation-only. Claude answers when you ask; it doesn’t sit on your account correlating alerts and acting on them. Close the chat and nothing continues.
- Laptop-bound for anything scheduled. The closest thing to “running on its own” is Claude Cowork’s scheduled tasks, which fire on a fixed clock, not on inbox events. That’s not an always-on, event-driven observability agent.
So Claude is great for “dig into this incident with me right now” and not built for “watch the fleet and act the moment something breaks.”
If you want New Relic work that runs on its own: Carly
The moment you want something to happen around New Relic without you in the chat — summarize an incident the instant an alert fires, post a daily reliability digest, chase the right owner when a service degrades — you’ve crossed past what Claude’s MCP server is for.
That’s where Carly fits. Carly is an AI executive assistant built to act on triggers, not just answer in a chat:
- Fires on events, 24/7, in the cloud — when a New Relic webhook fires, Carly acts; your laptop doesn’t need to be awake.
- Ties observability to the rest of your stack — turn an alert into a Slack message, an email to the owner, a Jira ticket, or a calendar hold for a postmortem, all in one flow.
- Actually sends and updates — drafts and sends email (Gmail and Outlook) with attachments, files and labels, manages tasks, updates your CRM, and records meetings.
- Builds the workflow for you — tell it “I’d like a system that summarizes every high-priority incident and emails the on-call” in plain English; it interviews you, then builds it with you. No prompt engineering.
AI agents start at $35/month, and steps in a workflow that don’t use AI run free and unlimited. Carly connects to 200+ tools across 40+ categories — see integrations. And Carly natively integrates with New Relic.
Claude’s New Relic MCP vs Carly
| Claude (New Relic MCP) | Carly | |
|---|---|---|
| Run NRQL queries | Yes | Yes |
| Inspect entities & dashboards | Yes | Yes |
| Read alerts & incidents | Yes | Yes |
| Acts on triggers / events | No | Yes |
| Summarizes incidents on its own | No | Yes |
| Works while laptop is closed | No | Yes (cloud) |
| Sends email as part of the flow | No (Gmail draft-only) | Yes (Gmail + Outlook) |
| Pricing | Pro $20 / Max $100–$200 | AI agents from $35/mo |
Claude’s MCP server is a strong New Relic query surface inside a chat. Carly is a teammate that acts on incidents as they arrive.
Frequently Asked Questions
Does Claude work with New Relic?
Yes. New Relic ships an official, hosted MCP server at https://mcp.newrelic.com/mcp/ (and an EU endpoint) that you add to Claude as a custom connector. Once connected, Claude can run NRQL queries and inspect entities, dashboards, alerts, and incidents inside a chat. It’s not a one-click directory connector, and like every MCP setup it only works in a conversation you start.
Can Claude monitor New Relic and act on alerts automatically?
No. The MCP server works inside a conversation you start — there are no event triggers, so Claude won’t summarize an incident or page an owner on its own. For automatic, trigger-based observability work, you need an agent platform like Carly.
How do I connect Claude to New Relic?
Get a scoped New Relic API key (or use OAuth), then in Claude go to Settings → Connectors → Add custom connector and point it at https://mcp.newrelic.com/mcp/. Authorize, optionally filter tools with include-tags headers, and query your telemetry in a normal chat. Remote custom connectors on claude.ai require a paid plan.
Is the New Relic MCP server GA?
New Relic broadly rolled it out through early 2026, but its docs still carry preview/pre-release language on parts of the server, so treat it as maturing rather than fully locked-down. There’s no separate pricing for the MCP server itself.
More: Claude connectors · Claude + Datadog · Claude + PagerDuty · Claude + Sentry · Claude vs Carly · Best AI agents for productivity
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
