- TEXT
- JSON
- CHAT
- HUMAN
Short answer
How can I get my Pulumi stack outputs in chat?
Your agent uses the get_stack_outputs tool to fetch specific values like IP addresses or URLs directly from your Pulumi stacks. You just ask for the specific key you need, and the data appears in your chat. This breakdown shows how the data moves from your stack to your conversation.
The results.
Where the data lands.
The specific information you request ends up in these formats.
TEXT
Direct values
The agent pulls raw strings like connection URLs or IP addresses. You get the exact data needed for your next command.
JSON
Structured data
If you ask for all outputs, the agent retrieves the full object. This is useful for debugging or passing data to other tools.
CHAT
Contextual answers
The agent interprets the output to answer your questions. You get a plain English explanation of what the output represents.
HUMAN
Manual verification
You remain in control of the infrastructure. The agent only reads the outputs and cannot change your stack configuration.
The workflow
What your AI does when you ask for outputs.
The agent acts as a bridge between your Pulumi state and your chat interface.
Identify the stack
The agent looks up the specific Pulumi stack you mentioned in your prompt.
get_stack_outputsFetch the values
The agent calls the tool to pull the requested keys from the Pulumi backend.
get_stack_outputsFilter the data
If you only asked for one specific URL, the agent ignores the rest of the stack data.
get_stack_outputsDeliver the result
The agent presents the connection string or IP address in your chat window.
get_stack_outputs
Try it
Copy these to start.
Use these prompts to get the data you need.
Starting points
Swap out the stack names and key names for your actual Pulumi resources.
Pulumi Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_v2N5nlYKKcBMg7t6tvswP5zOuY1KsfB14QgLHciO/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Pulumi capabilities are ready to use.
{
"mcpServers": {
"pulumi-mcp": {
"url": "https://edge.vinkius.com/vk_preview_v2N5nlYKKcBMg7t6tvswP5zOuY1KsfB14QgLHciO/mcp"
}
}
}What is the connection string for the 'production' stack?
Give me the IP address for the load balancer in my 'staging' stack.
List all the outputs from my 'dev-cluster' stack.
Find the endpoint URL for the database in the 'main' stack.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Pulumi once, then ask.
Click the link to connect your Pulumi account. Your credentials stay encrypted, and you can immediately start asking your agent for stack data.
Connect Pulumi to your AIFAQ
How this task behaves.
- 01
Can the AI change my Pulumi stack settings?
No. This MCP only has the ability to read outputs. It cannot modify, delete, or create infrastructure.
- 02
Do I need to run Pulumi commands in my terminal?
No. Once connected, you can get these values entirely through your AI client.
- 03
Can the AI see my entire Pulumi configuration?
No. It only sees the specific outputs you ask for or the list of outputs provided by the stack.
- 04
What happens if a stack output is a secret?
The agent can only retrieve outputs that are explicitly marked as visible or accessible via the tool.
- 05
Can I use this to update my infrastructure?
No. This specific tool is for retrieving data, not for running updates or deployments.
More questions about Pulumi? The Connector page answers them. See everything the Pulumi Connector can do
Connect Pulumi to Claude, Cursor, ChatGPT & more
