Manage incidents from

Claude Code, Cursor, or VS Code

Your AI agent already knows your codebase. Now it can create incidents, page on-call engineers, and track resolution through Runframe. No new tool to learn.

View on GitHubRead the full case

Your agent already knows your system better than ours ever will.

The agent running in your IDE has read your codebase, seen your commits, and helped you debug at 2 AM. It knows your deploy history and your architecture. It just can't create an incident, page someone, or update a timeline. Runframe fixes that.

What your agent can do with Runframe

Tools across the full incident lifecycle.

One-line MCP setup

npx @runframe/mcp-server --setup. Works with Claude, Cursor, Windsurf, VS Code. One config file, no custom integration.

Create and manage incidents

Declare incidents with severity, change status, acknowledge, update, and resolve. All through structured MCP calls.

Page on-call engineers

When the agent detects something it can't fix alone, it pages the on-call engineer. Slack DM, email, and escalation policies all apply.

Write to the timeline

Agents add investigation findings, log links, and remediation steps to the incident timeline. The human responder walks in with full context.

Check on-call and escalation

Look up who is on-call for a service, check escalation policies, list teams. The agent knows who to page before it pages them.

Create postmortems

After resolution, agents can create postmortem drafts from incident data. Timeline, impact, and action items are already captured.

How it works

1

Install the MCP server

npx @runframe/mcp-server --setup. It writes the config file for your AI tool. Takes 30 seconds.

2

Agent detects an issue

Your monitoring agent, CI pipeline, or coding assistant spots a problem. It creates an incident via MCP with severity and context.

3

On-call is paged

Runframe pages the on-call engineer via Slack and email. The agent's findings are already on the timeline.

Captive vs. open

Vendor-built agentYour agent + Runframe

ContextVendor data onlyYour codebase + infra

ModelVendor's choiceYour choice

PortabilityLocked to vendorWorks across tools

CustomizationVendor featuresYour workflows

Frequently asked questions

What is MCP?

Which tools work with the MCP server?

What can an agent actually do through the MCP server?

Can agents page humans?

Do agent-created incidents show up in analytics?

What if I do not use AI agents yet?

Is this different from just having an API?

What about data privacy?

Your agent. Your models. Our incident API.

Start free. Connect your first agent in minutes.

Start FreeView All Integrations