← Back to Home

Connect OpenClaw to Your Data

A step-by-step guide to set up Expanso pipelines and wire them into OpenClaw. Your credentials stay on your machine.

⚠️

Expanso Cloud Account Required

You'll need a free Expanso Cloud account to use expanso-edge. The CLI authenticates with Cloud to validate pipelines and manage deployments. Create your free account →

Prerequisites

This tutorial assumes you have OpenClaw installed and running. If not, get OpenClaw first.

Create Your Expanso Cloud Account

Expanso Cloud is required to use expanso-edge. It handles authentication, pipeline validation, and deployment orchestration. The free tier is generous—no credit card needed.

1.1 Sign up for free

Create your account at cloud.expanso.io:

# Open the signup page in your browser open https://cloud.expanso.io/signup # Or just visit: cloud.expanso.io/signup

1.2 Install the Expanso CLI

Once you have an account, install the CLI tools:

# macOS / Linux - install expanso-edge curl -fsSL https://get.expanso.io/edge/install.sh | bash # Verify installation expanso-edge --version # expanso-edge v1.x.x

1.3 Connect your machine to Cloud

Get your bootstrap token from the Cloud dashboard, then authenticate:

# Authenticate expanso-edge with your Cloud account expanso-edge bootstrap --token "exso_bt_your_token_here" # ✓ Connected to Expanso Cloud # ✓ Node registered: your-machine-name
Why Cloud? Expanso Cloud provides pipeline validation, secret management, scheduling, monitoring, and team collaboration. Your credentials and data still stay local—Cloud just orchestrates.
Account ready? Let's build a pipeline

Build and Validate a Pipeline

Create your first pipeline and validate it against Expanso Cloud before running.

2.1 Create a simple pipeline

Save this as hello-pipeline.yaml:

input: generate: count: 1 mapping: | root.message = "Hello from Expanso!" root.timestamp = now() pipeline: processors: - mapping: | root = this root.processed = true output: stdout: {}

2.2 Validate with Cloud

Check your pipeline syntax before running:

# Validate pipeline syntax (offline, no auth required) expanso-cli job validate hello-pipeline.yaml --offline # ✓ Pipeline syntax valid

2.3 Run locally

Test your pipeline on your machine:

# Run the pipeline locally expanso-edge run --config hello-pipeline.yaml # {"message":"Hello from Expanso!","timestamp":"2025-...","processed":true}
Browse ready-made pipelines: We have 172 pre-built pipelines for common tasks like email triage, Stripe reports, and DevOps monitoring. Browse the catalog →
Pipeline working? Let's deploy it

Deploy to Production

Deploy your pipeline with scheduling, monitoring, and team access via Expanso Cloud.

3.1 Deploy your pipeline

Deploy your pipeline to run on your edge node:

# Deploy the pipeline expanso-cli job deploy hello-pipeline.yaml # ✓ Pipeline deployed: hello-pipeline # ✓ Job ID: job-abc123

3.2 Monitor your jobs

Check status and view logs:

# List running jobs expanso-cli job list # Check specific job status expanso-cli job describe job-abc123 # View job logs expanso-cli job logs job-abc123
Secrets stay local: Your API keys (OpenAI, Stripe, etc.) are stored in environment variables on your edge node. Cloud never sees them—it only sends pipeline configs.
Final step: connect to OpenClaw

Connect to Your AI Assistant

Add the Expanso MCP server to OpenClaw, Claude Desktop, or any MCP-compatible client.

4.1 For OpenClaw

Register the Expanso MCP server:

expanso-cli mcp add npx @expanso/mcp-pipelines

4.2 For Claude Desktop

Add to your claude_desktop_config.json:

{ "mcpServers": { "expanso-pipelines": { "command": "npx", "args": ["@expanso/mcp-pipelines"], "env": { "EXPANSO_API_KEY": "your-api-key-here" } } } }

4.3 Available MCP Tools

Once connected, your AI assistant can use these tools:

validate_pipeline → Validate pipeline YAML create_pipeline → Generate pipeline from template deploy_pipeline → Deploy to Expanso Cloud list_pipelines → List deployed pipelines list_skills → Browse 170+ pre-built skills
Security: Your API keys (OpenAI, Slack, Stripe, etc.) are stored in environment variables on your machine. The MCP server reads them locally—they're never sent to the AI.

You're All Set! 🎉

Browse our catalog of 172 ready-to-use pipelines, or build your own.

Need Help?