What teams do with Escanor

All use cases

• Incidents

  AI incident response

  Gather alerts from connected providers, investigate with the evidence attached, and approve a bounded repair or rollback before it ships.

  • Alerts from Sentry, PagerDuty, GitHub and Vercel
  • Evidence linked to recent deploys and logs
  • Repairs wait for an owner or manager
  Read the use case
• Coding agents

  Safe deploys for AI coding agents

  Give coding agents one revocable MCP key, require confirmation for destructive operations, and keep redeploys and rollbacks behind human approval.

  • Works with Claude Code, Cursor, VS Code and ChatGPT
  • Destructive calls need confirm=true
  • Rollbacks start as a dry run
  Read the use case
• Machines

  Run Claude Code on your own servers

  Connect a Linux server with an outbound agent, chat with Claude Code sessions on it from anywhere, and approve each command from your phone.

  • No inbound ports to open
  • One Claude Code session per chat
  • Allow or deny each command
  Read the use case
• Phone

  Manage Kubernetes from your phone

  Run Claude Code on a server you own, check your cluster from the Android or iPhone app, and allow or deny each command from a card on your phone.

  • Your kubeconfig stays on your server
  • Android and iPhone apps
  • Approve every write
  Read the use case
• Automation

  Automate operations with Autopilot

  Give Autopilot a goal and “Done when” criteria, run it on triggers, and decide on anything risky under Needs you, within budgets you set.

  • Goals, triggers and multi-step missions
  • Autonomy levels and budgets
  • Stop everything now pauses it all
  Read the use case
• Monitoring

  Monitor production in one place

  See deployments, incidents, logs and activity from every connected tool side by side, and send your own logs to Escanor with an ingest key.

  • Deploys, incidents and logs together
  • One activity timeline across tools
  • Custom logs through an ingest key
  Read the use case

How Escanor helps

Every use case is built from the same pieces, so what you set up for one carries over to the next.

Connect your stack

Code hosts, clouds, databases, monitoring and chat from a catalog of hundreds of providers.

Learn more

Give agents one key

Claude Code, Cursor, VS Code and ChatGPT reach every connected tool through one MCP server.

Learn more

Investigate incidents

Alerts, deploys and logs in one place, with an assistant that links its evidence.

Learn more

Automate with approval

Autopilot runs goals and triggers, and risky steps wait for an owner or manager.

Learn more

Drive your servers

Chat with Claude Code on machines you own and approve each command.

Learn more

Work from anywhere

The web dashboard, a desktop app with local models, and Android and iPhone apps.

Learn more

Start in three steps

1. Step 1

   Connect one provider

   Pick a tool you already use, such as GitHub or Vercel, from the integrations catalog.

2. Step 2

   Run a safe read

   Ask the assistant to list your projects or recent deployments, and check the answer.

3. Step 3

   Approve your first change

   Let an agent propose one bounded change, then approve it and verify the result.

The quickstart walks through each step.