Here's a scenario every on-call engineer knows too well: a critical incident fires for a service you’ve never seen before. Your first ten minutes are a frantic scramble across wikis and Slack channels just to answer the most basic questions: Who owns this? What does it do? Where are the runbooks? By the time you’re oriented, the incident has escalated.
In response, we built the Cortex Model Context Protocol (MCP) to bring the context of your entire engineering ecosystem directly into the tools your team already uses. After a successful research preview, we’re excited to announce that the MCP is now generally available.
With expanded API support and new data integrations, the MCP allows you to get the full benefits of Cortex right where you work. Here's what that means for your developers and engineering leaders.
The engineering knowledge you need, on-demand
Your Cortex catalogs are the single source of truth for your engineering ecosystem. But until now, that context has been trapped in a web UI. Day-to-day development and incident response happen elsewhere, increasingly inside LLMs like Claude and ChatGPT. The MCP is designed to eliminate this costly context switching.
It connects your favorite chat interface directly to your Cortex catalog, giving any developer immediate access to the ownership information, documentation, and operational data they need to make decisions and take action, without ever leaving their workflow.
How the MCP helps across your engineering organization
The power of the Cortex MCP is how it adapts to the specific needs of different roles across your team. While we’re only scratching the surface, here are a few ways the Cortex MCP can be a massive unlock across your entire engineering organization.
For engineers and SREs on the front lines
MCP helps developers and SREs get the information they need without leaving their IDEs or chat clients. From critical context during incidents to proactively fixing gaps, the answers are always on-demand.
Let’s revisit the on-call scenario. An incident fires for an unfamiliar service. Instead of a frantic ten-minute scramble, the engineer asks their chat client a simple prompt:
“Who’s on call for orders-api right now and when was it last deployed?”
They instantly get the owner, the runbook, and a list of recent changes. This information allows them to begin an investigation immediately armed with the right information.
For platform teams
For platform teams, the MCP is a powerful tool to operationalize standards and drive strategic initiatives. It allows them to move from passive tracking to active management of their platform goals. Instead of chasing down teams in Slack, a platform engineer can get a clear picture of progress with a single question:
“Which services are blocking completion of the modernization Initiative?”
This allows them to spot adoption gaps, identify teams that are falling behind, and keep key initiatives moving forward.
For engineering leaders
The Cortex MCP gives leaders real-time insights into team performance and compliance, enabling them to make better decisions without waiting for reports. Instead of synthesizing data from multiple systems, a VP of Engineering can ask their LLM directly to understand the business impact of their investments:
“How has AI adoption impacted cycle time and incidents in the last quarter?”
By grounding the AI in the same single source of truth that developers use, the MCP gives leaders the ability to query service health and track Scorecard progress, turning your catalog into an intelligent decision-making engine.
Solving daily challenges with the Cortex MCP
The power of the Cortex MCP is how it adapts to different needs. For a new developer, it’s an onboarding assistant. For a manager, it’s an executive dashboard. For an on-call engineer, it’s a critical incident co-pilot.
Imagine a developer joining a new team. Instead of spending a week navigating wikis, they can simply ask their LLM, "What are the most important services for the Checkout team, and are there any high-priority issues I should know about?" They get up to speed in minutes, not days.
This same power transforms high-stakes situations. Let’s revisit the on-call scenario, this time with the MCP. You're in Claude, investigating an active incident for an unfamiliar service.
You start with a simple prompt:
"Use the Cortex MCP to tell me about the
cortex-brainservice."
Instantly, you get the critical context: the service owners, the team’s Slack channel, a high-level description, the escalation contact, and its service tier. You have points of contact and documentation links in seconds.
You dig deeper with a follow-up:
"Show me recent deploys for
cortex-brain."
You now see a list of the most recent deploys and deployment patterns, giving you a list of recent changes that could have impacted the service. You can even ask, "Have there been recent incidents?" to check the service's history.
In under two minutes, without leaving your chat window, you’ve gathered the critical context that used to take ten minutes of frantic searching. Your investigation can start immediately, armed with the right information.
The Cortex MCP is a new way for your teams to access the knowledge trapped across an entire ecosystem of apps, documents, and people. By bringing your engineering ecosystem into the conversational tools your teams love, we’re making it easier than ever to build, own, and operate software with confidence.
Ready to bring the power of your Cortex catalog to your favorite LLM? Get started with the MCP today.
