Skip to main content
Back to the Pulumi MCP server

MCP Gateway · Claude Code

Pulumi icon

Connect Pulumi to Claude Code

Writing infrastructure code means checking provider schemas constantly. Claude Code can look up Pulumi Registry resources and types while it edits your program, then search what is already running before it adds more. Speakeasy's hooks check each prompt and tool call before Claude Code runs it, so a deploy can be denied before it starts.

Category
Infrastructure
Tools
14
Authentication
OAuth
Claude Code
CLI

Behind the MCP Gateway

Run Pulumi MCP behind the MCP Gateway

One URL for every agent, sign-in through your IdP, RBAC down to the tool, and every Pulumi call inspected and logged.

Deploy on MCP Gateway

Set up

Add Pulumi to Claude Code in 4 steps

  1. 01

    Add Pulumi from the catalog

    An admin finds Pulumi in the MCP Catalog (MCP Gateway > MCP > Catalog) and clicks Add. Speakeasy registers the vendor's server as a remote MCP server and proxies sessions to it. Nothing is copied or re-hosted.

    Pulumi uses OAuth. Servers that support dynamic client registration get their OAuth configuration applied automatically.

    Where MCP-scoped policies are enabled, the install offers a recommended risk policy for the new server that detects secrets and PII and inspects tool requests and responses.

  2. 02

    Choose who can use Pulumi

    Give teams the mcp:connect scope on the Pulumi server through team access or a role. The organization gateway works out each person's servers from their live access, so granting access adds Pulumi to their gateway on the next call, and removing it takes it away.

  3. 03

    Connect Claude Code

    Add the organization gateway as an HTTP MCP server, then authenticate from /mcp.

    Run the command below with your organization gateway URL.

    Open /mcp in Claude Code and choose Authenticate to sign in.

    claude mcp add --transport http --scope user \
      'speakeasy' \
      'https://<organization host>/mcp/gateway/<organization slug>'
  4. 04

    Sign in and use Pulumi from Claude Code

    The first time Claude Code connects, it opens a browser so the person can sign in to Speakeasy through your identity provider.

    Pulumi needs its own OAuth grant, so each person connects it once. Until they do, Pulumi calls fail while the rest of the gateway keeps working.

    From then on, Pulumi's tools appear in Claude Code, limited to what the person's access allows, and every call lands in the gateway's tool logs.

To roll it out to a whole team, publish the server in a plugin and register the plugin marketplace in Claude Code's managed settings, so every member gets it without running the command. The device agent can also write and reapply these settings on each laptop.

Use cases

What Claude Code users do with Pulumi

Look up a resource's inputs while coding

When Claude Code adds a bucket or a queue to your program, it lists the provider module's resource types and reads the exact inputs and nested types before writing them.

list-resourcesget-resourceget-type

Check what already exists in production

Before creating a resource, search the Pulumi-managed resources in your production stack for one with the same name or type, so you reuse it instead of duplicating it.

resource-search

Deploy a prototype to AWS

For a quick test environment, Claude Code analyzes the application files in your checkout and generates Pulumi infrastructure that provisions the AWS resources they need.

deploy-to-aws

Tools

14 tools

Each Pulumi tool can be allowed or denied per team and role at the MCP Gateway before Claude Code sees it.

get-stacks

Call this tool to list Pulumi stacks for the authenticated user and organization. , "stacks with 'prod' in the name", "stacks created after 2024", "show me stacks containing 'bob'"). Use resource-search instead for these filtered queries.

get-policy-violations

Get policy violations for Pulumi stacks. Use this tool to answer questions like: - "Do I have any policy violations?" - "What security issues are in my infrastructure?" - "Show me policy violations for my production stack" - "Are there any compliance issues in my project?"

get-users

List all users (members) in the organization. Use this when asked about users, members, admins, or team members.

neo-bridge

Launch and monitor Neo tasks step by step. Pulumi Neo is a purpose-built cloud infrastructure automation agent. If the JSON result has has_more=true, call this tool again to read more data. Continue calling until has_more=false.

neo-reset-conversation

Reset the Neo conversation for a specific task

See all 14 tools

Control

What Speakeasy can see and stop in Claude Code

The gateway covers every client the same way. The rest depends on what Claude Code exposes to Speakeasy.

Tool logs

Gateway

Every Pulumi call is logged with the person and the client that made it.

Access by role

Gateway

People reach only the Pulumi tools their team or role allows.

Request and response blocking

Gateway

A risk policy scoped to Pulumi can reject a tool request before it reaches the server, or withhold a result from Claude Code. MCP-scoped policies are enabled per organization.

Prompt and tool-call blocking

Claude Code

Agent hooks send each prompt and tool call to Speakeasy, which allows or denies it before Claude Code runs it.

User attribution

Claude Code

Claude Code's OpenTelemetry exporter ties each session to a named person. Without it, activity can't be attributed.

Budgets

Claude Code

Spend limits can block Claude Code traffic once a budget is reached.

Device agent and MDM

Claude Code

The device agent installs the marketplace, plugin and telemetry settings and reapplies them every minute. It deploys through MDM.

Anthropic inference hooks

Claude Code

On Claude Enterprise, prompts can be checked before inference with nothing installed on the device (beta).

  • Claude Code on the web doesn't use the laptop setup. It needs the device agent in a shared Anthropic-hosted environment.
  • Plugin marketplace URLs for Claude Code carry an access token, so share install snippets only through a private channel.

Questions

Frequently asked questions

Can I stop Claude Code from deploying with Pulumi?
Yes. You decide which of the 14 Pulumi tools each team can call at the MCP Gateway, so a team can read the registry and stacks without deploy-to-aws. Agent hooks also allow or deny each tool call before Claude Code runs it.
Does Claude Code hold a Pulumi access token?
No. Sessions authenticate with OAuth 2.1 through your identity provider, and agents never hold a Pulumi credential. Access ends when the person leaves the directory.
How do Claude Code users get Pulumi?
An admin adds Pulumi from the Speakeasy MCP Catalog and grants access by team or role. Add the organization gateway as an HTTP MCP server, then authenticate from /mcp. Each person then sees only the Pulumi tools their access allows.
Are Pulumi calls from Claude Code logged?
Yes. Pulumi calls go through the MCP Gateway, which logs each tool call and attributes it to the person and the client that made it.

Connect more

Pulumi in other clients

AI everywhere.

Control here.