Skip to main content

Overview

Qovery AI Copilot is your intelligent assistant for managing infrastructure. It’s available through three interfaces: the Console (built-in), Slack Bot (team collaboration), and MCP Server (power users). All allow you to interact with your Qovery infrastructure using natural language.

Operating Modes

The Qovery AI Copilot operates in two modes:

Read-Only Mode (Default)

Safe exploration and information access
  • View information and status
  • Get help and recommendations
  • Troubleshoot issues
  • Learn about features
  • Cannot make changes to infrastructure
Available in: Console (only), Slack Bot, MCP Server

Read-Write Mode

Full automation capabilities
  • Everything in read-only mode
  • Deploy and restart services
  • Create/modify/delete resources
  • Update configurations
  • Manage environment variables
Available in: Slack Bot, MCP Server (not Console)
Default Mode: By default, the Copilot operates in read-only mode for safety. The Console Copilot is permanently in read-only mode. Slack Bot and MCP Server support read-write mode when needed.
Early Access Required: Qovery AI Copilot is available online and on demand. Contact Qovery Support to request early access.
Beta Status: Qovery AI Copilot is currently in beta. Features and capabilities are being actively developed.

Choose Your Interface

Qovery AI Copilot is available through three interfaces:

Prerequisites

  • Qovery Account: Active account with infrastructure deployed
  • For Console: No additional setup needed! Just log into the Qovery Console
  • For Slack Bot: Slack workspace and API token
  • For MCP Server: Claude Desktop and API token

Quick Start

Choose your preferred interface to get started:
  • Console (Easiest)
  • Slack Bot (For Teams)
  • MCP Server (For Power Users)
1

Log into Console

2

Click AI Copilot

Click the AI Copilot button (top-right) or press ⌘ I
3

Start Asking

Ask questions or click on suggested actions
Full Console Copilot Guide →

Your First Interactions

Check Infrastructure Status

"Show me all my environments"
"What's the status of production?"
"List all applications in staging environment"

Deploy Applications

"Deploy the latest version of my-api to staging"
"Redeploy the frontend application"
"Stop all services in the dev environment"

Investigate Issues

"Why is my application deployment failing?"
"Show me error logs from the last hour"
"Check if the database is running"

Understanding How Copilot Works

Natural Language Understanding

Qovery AI Copilot understands natural language requests:
  • Casual Language
  • Technical Language
  • Mixed Approach
✅ “Can you deploy my app?” ✅ “Stop everything in dev” ✅ “Show me what’s broken”

Context Awareness

Copilot maintains conversation context:
You: "Show me the production environment"
Copilot: [Displays production details]

You: "Clone it to staging-test"
Copilot: [Creates clone without asking which environment]

You: "Now deploy version 2.0 there"
Copilot: [Deploys to staging-test automatically]

Autonomous Execution

For complex tasks, Copilot breaks them down and executes automatically: Your Request: “Create a staging environment for the new feature” Copilot’s Actions:
  1. Creates new environment
  2. Configures networking
  3. Sets up databases
  4. Deploys applications
  5. Validates everything is running
  6. Reports success

Best Practices

Be Specific When Needed

  • Good
  • Can Be Improved
✅ “Deploy my-api to production” ✅ “Scale the database in staging” ✅ “Show logs for the last 30 minutes”
If Copilot needs more information, it will ask clarifying questions.

Use Follow-Up Questions

Build on previous interactions:
You: "Show me production resources"
Copilot: [Shows resources]

You: "Which ones are using the most CPU?"
Copilot: [Analyzes and reports]

You: "Scale up the top one by 2x"
Copilot: [Scales the identified service]

Leverage Agentic Capabilities

Let Copilot handle complex workflows: Instead of:
1. "Show me all environments"
2. "Check which ones are inactive"
3. "Calculate cost savings"
4. "Stop the inactive ones"
Just ask:
"Stop all environments that have been inactive for more than 24 hours"
Copilot will plan and execute all steps autonomously.

Safety & Guardrails

Confirmation for Destructive Actions

Copilot will ask for confirmation before:
  • Deleting environments or services
  • Stopping production workloads
  • Making infrastructure changes

Respects Permissions

Copilot can only perform actions you’re authorized to do:
  • Uses your Qovery credentials
  • Follows RBAC policies
  • Cannot bypass security controls

Audit Trail

All Copilot actions are logged:
  • View in Qovery Console audit logs
  • Track what was changed and when
  • Identify which actions Copilot performed

Common Use Cases

Daily Operations

"Show me the status of all services"
"Deploy the latest commit to staging"
"Restart the API service"
"Check database connection"

Troubleshooting

"Why is my deployment failing?"
"Investigate errors in production"
"Show me logs for pod xyz123"
"Debug the connection timeout"

Cost Optimization

"Show me which services cost the most"
"Find underutilized resources"
"Stop non-production environments for the weekend"
"Analyze spending over the last month"

Infrastructure Management

"Create a new staging environment"
"Clone production to a test environment"
"Update environment variables for all services"
"Scale up the database"

Getting Help

If you’re unsure what Copilot can do:
"What can you help me with?"
"Show me examples of what you can do"
"Help me troubleshoot an issue"
Copilot will provide guidance and suggestions.

Limitations (Beta)

Current limitations while in beta:
  • Console is Read-Only: Console Copilot operates permanently in read-only mode for safety
  • Default Mode is Read-Only: Slack Bot and MCP Server default to read-only mode
  • Confirmation Required: Destructive operations require user confirmation in read-write mode
  • Beta Features: Some capabilities still being developed
  • Response Time: Complex operations may take time

Next Steps

Feedback & Support

We’re actively improving Qovery AI Copilot. Share your feedback: Your feedback helps us build better automation!