Remote MCP for production systems

Investigate production from your AI assistant.

Connect ChatGPT, Codex and Claude to Kubernetes, observability and incident context through one secure remote MCP server.

Connect IncidentFlowWatch a 60-second demo
Read the docs

Remote MCP · Production integrations · Browser-based setup

02 / ALERT DETECTED

A production signal changes.

03 / INVESTIGATION STARTS

Your AI assistant starts an investigation.

04 / EVIDENCE GATHERED

IncidentFlow gathers evidence across your stack.

ChatGPT
Codex
Claude
IncidentFlowREMOTE MCP
Kubernetes
Grafana
Prometheus
Loki
Slack
GitHub
Argo CD
IncidentFlow MCPConnected
remote_mcp_ready
Waiting for a production signal.

Works with the tools already running in production

  • Kubernetes
  • Prometheus
  • Grafana
  • Loki
  • Argo CD
  • GitHub
  • Slack

Quickstart

Connect IncidentFlow in one command.

Add the remote MCP server to Codex or Claude Code, authenticate in your browser, and start investigating production.

Ready in under a minute.

codex mcp add incidentflow --url https://mcp.incidentflow.io/mcp

Then verify: codex mcp list

No local MCP process to maintain.

  1. 01
    Add MCP server

    Use the command for your AI client.

  2. 02
    Authenticate

    Authorize IncidentFlow in your browser.

  3. 03
    Investigate

    Ask for an explanation of production changes.

First prompt

Check the production cluster for unhealthy workloads and explain what changed.
MCP connectedWorkspace resolvedKubernetes available

Product

See the investigation happen.

IncidentFlow turns production signals into a traceable sequence of MCP tool calls, evidence and next steps.

Connect in under a minute

Codex adds IncidentFlow as a remote MCP server and confirms that the connection is ready.

Integrations

One MCP connection. Your production context.

Your AI assistant connects through IncidentFlow remote MCP to the infrastructure, telemetry and context that explain an incident.

AI assistants

Where the investigation begins.

ChatGPTAsk from the interface your team already uses.
CodexInvestigate from a developer workflow.
ClaudeBring production context into an AI conversation.

Infrastructure

Workload and deployment context.

KubernetesInspect unhealthy workloads and logs.
Argo CDUnderstand deployment state.
GitHubConnect code changes to incidents.

Observability

Metrics, logs and operational evidence.

GrafanaRead dashboard context.
PrometheusQuery production metrics.
LokiSearch relevant logs.

Team context

The operational conversation around an incident.

SlackFind incident and deployment context.
View integration docs

Security

Production access should be explicit.

Access path

UserAuthorizationWorkspaceIntegration

IncidentFlow is a remote MCP server for production investigations. Connection and integration access are intentionally presented as explicit operational choices.

  • Remote MCP endpoint for centralized operation
  • Integration connections are configured deliberately
  • Investigation context stays traceable through tool calls
  • Security details live in a dedicated overview
Read security overviewHow authentication works

Start investigating

Bring production context to your AI assistant.

Connect IncidentFlow through remote MCP and start your first investigation.

Connect IncidentFlowRead the docs
codex mcp add incidentflow --url https://mcp.incidentflow.io/mcp

Remote MCP endpoint ready for Codex