Knowledge

Knowledge is a persistent, versioned document store that gives your AI agent the context it needs to act accurately on your infrastructure.

What Knowledge Does

The AI agent knows your infrastructure resources and change history from connected integrations. Knowledge extends that with context only your team can provide:

• Runbooks and incident playbooks
• Architecture decisions and service ownership
• On-call procedures and escalation paths
• Notes about how your systems actually behave
• Any context that cannot be discovered automatically

Files you add to Knowledge are loaded into the AI agent's context, so it can reference them when answering questions, investigating incidents, and executing workflows.

Creating a File

1. Navigate to Knowledge in the left sidebar
2. Click New File
3. Give the file a name and optionally organize it into a folder using path notation (e.g. team-notes/deploy-rollback)
4. Write your content in the markdown editor
5. Changes are saved automatically

Files support standard markdown — headings, lists, code blocks, tables, and links.

Organizing with Folders

Use / in the file name to create a folder structure. Folder names can be whatever works best for your team:

team-notes/deploy-rollback
incidents/database-failover
platform/service-map
ops/escalation-policy

Folders appear in the sidebar and keep your knowledge base navigable as it grows.

Version History

Every save creates a new version. You can view the full history of any file and restore a previous version at any time.

1. Open a file
2. Click the History icon in the toolbar
3. Select any past version to preview it
4. Click Restore to roll back to that version

Up to 50 versions are retained per file.

How the AI Agent Uses Knowledge

When you send a message in chat, the AI agent pulls relevant Knowledge files into its context alongside your live infrastructure data. This means:

• It can reference your runbooks when investigating an incident
• It can follow your documented procedures when executing workflows
• It can surface ownership and escalation information from your notes
• Its answers are grounded in how your team has described the system, not just what the APIs report

The more context you add to Knowledge, the more accurately the agent can reason about your specific environment.

What to Add

Good candidates for Knowledge files:

• Runbooks — step-by-step procedures for common operational tasks
• Architecture notes — descriptions of how services connect and depend on each other
• Service ownership — who owns what, who to contact when something breaks
• Incident history — summaries of past incidents and how they were resolved
• Deployment notes — quirks and requirements for specific services
• On-call guides — what to check first, common failure patterns, escalation paths

Limits

For questions about storage limits by plan, contact support.