MCP server

The walkerOS MCP server (@walkeros/mcp) exposes CLI tools and the walkerOS API as Model Context Protocol tools. This lets AI assistants like Claude validate events, bundle flows, simulate pipelines, and manage projects without leaving the conversation.

Quick start

Add the server to your Claude Desktop or Claude Code configuration:

Loading...

The server starts on STDIO and registers all tools automatically.

Environment variables

Variable	Required	Default	Purpose
WALKEROS_TOKEN	Yes (API tools)	—	Bearer token (sk-walkeros-...)
WALKEROS_PROJECT_ID	Yes (project-scoped tools)	—	Active project ID (proj_...)
WALKEROS_APP_URL	No	https://app.walkeros.io	Base URL override

CLI tools (validate, bundle, simulate, push) work without any environment variables. API tools (whoami, flows, projects, versions) require WALKEROS_TOKEN.

Available tools (20)

Local CLI tools

These wrap the walkerOS CLI for use by AI assistants.

validate

Validate walkerOS events, flow configurations, or mapping rules.

• type (required): "event" | "flow" | "mapping"
• input (required): JSON string or file path to validate
• flow (optional): Flow name to validate against

bundle

Bundle a walkerOS flow configuration into deployable JavaScript.

• configPath (required): Path to the flow configuration file
• flow (optional): Specific flow name to bundle
• stats (optional): Include bundle statistics in output
• output (optional): Output file path for the bundle

simulate

Simulate events through a walkerOS flow without making real API calls. Processes events through the full pipeline including transformers and destinations.

• configPath (required): Path to the flow configuration file
• event (required): JSON string representing the event to simulate
• flow (optional): Flow name for multi-flow configs
• platform (optional): "web" | "server" (default: "web")

push

Push a real event through a walkerOS flow to actual destinations. This makes real API calls to real endpoints.

• configPath (required): Path to the flow configuration file
• event (required): JSON string representing the event to push
• flow (optional): Flow name for multi-flow configs
• platform (optional): "web" | "server" (default: "web")

API tools

These require WALKEROS_TOKEN to authenticate with the walkerOS API.

Auth

Tool	Purpose
whoami	Verify your API token and see your identity

Projects

Tool	Purpose
list-projects	List all projects you have access to
get-project	Get details for a project
create-project	Create a new project
update-project	Update a project name
delete-project	Soft-delete a project and all its flows

Flows

Tool	Purpose
list-flows	List all flow configurations in a project
get-flow	Get a flow with its full content (Flow.Setup JSON)
create-flow	Create a new flow configuration
update-flow	Update a flow name and/or content (creates version snapshot)
delete-flow	Soft-delete a flow configuration
duplicate-flow	Create a copy of an existing flow

Versions

Tool	Purpose
list-versions	List version history for a flow
get-version	Get a specific version of a flow
restore-version	Restore a flow to a previous version

Remote bundling

Tool	Purpose
bundle-remote	Bundle a flow using the walkerOS cloud service (no local build tools needed)

Example workflows

Validate and simulate a flow

Ask your AI assistant:

"Validate my flow config at ./walker.config.json, then simulate a page view event through it."

The assistant will call validate first, then simulate with the event — all without leaving the conversation.

Manage flows via the API

"List all flows in my project and show me the content of the most recent one."

The assistant calls list-flows, picks the latest, then calls get-flow to retrieve the full Flow.Setup JSON.

Bundle and deploy

"Bundle the default flow from ./walker.config.json and show me the bundle stats."

The assistant calls bundle with stats: true and returns the output.

Installation

Loading...

Peer dependencies

The MCP server requires the walkerOS CLI as a peer dependency:

Loading...

Next steps

• CLI — Learn about the underlying CLI commands
• Flow configuration — Understand flow structure
• MCP specification — Learn about the Model Context Protocol