Deploy your Claude Code agents to cloud.atxp.ai for pay-per-use execution.
- 🚀 One-command deployment - Deploy your agents with
/deploy - 🔐 Secure authentication - Uses your ATXP connection string
- ⚙️ Environment variables - Configure secrets and API keys with 600 permissions
- 🔄 Update existing instances - Seamlessly update deployed agents
- Bash 3.2+ (default on macOS and most Linux distributions)
- Git (for deployment tracking)
- Write access to project directory
First, configure your ATXP connection string:
/setup <your-connection-string>Get your connection string from your ATXP dashboard.
Deploy your current directory to the cloud:
/deployThis will:
- Create a zip of your project (excluding sensitive files)
- Upload to cloud.atxp.ai
- Return your instance URL
- Save the instance ID in
.atxp-instancefor future updates
If your agent needs API keys or secrets:
/configure-env-var GOOGLE_ANALYTICS_API_KEY ya29.a0AfH6SMBx...
/configure-env-var DATAFORSEO_API_KEY 12345678-1234...Then redeploy:
/deployConfigure your ATXP connection string for authentication.
/setup <connection-string>Deploy your agent to cloud.atxp.ai.
/deployWhat gets deployed:
- All project files and directories
.atxp/.env.production(if configured)
What's excluded:
.envfiles (except.atxp/.env.production)node_modules/.git/- Credentials (
.npmrc,.aws/*,*.pem, etc.) - See
/deploycommand for full exclusion list
Set an environment variable for your deployed agent.
/configure-env-var KEY VALUEExample:
/configure-env-var DATABASE_URL postgresql://user:pass@host:5432/dbVariables are stored in .atxp/.env.production and included in deployments.
View all configured environment variables (values masked).
/list-env-varsExample output:
Environment variables in .atxp/.env.production:
DATABASE_URL=post...****
API_KEY=sk-1...****
(2 variables configured)
Remove an environment variable.
/remove-env-var KEYExample:
/remove-env-var OLD_API_KEYUse environment variables for:
- API keys and secrets
- Database connection strings
- Third-party service credentials
- Configuration that differs between environments
- Variables are stored in
.atxp/.env.production - This file is included in deployments (unlike other
.envfiles) - Your deployed agent can access them at runtime
- The file is automatically added to
.gitignorefor security
# Configure multiple API keys
/configure-env-var GOOGLE_ANALYTICS_API_KEY ya29.a0AfH6SMBx...
/configure-env-var GOOGLE_SEARCH_CONSOLE_KEY AIzaSyC...
/configure-env-var DATAFORSEO_API_KEY 12345678-1234...
# Review what's configured
/list-env-vars
# Deploy with environment variables
/deploy
# Later: Update a variable
/configure-env-var DATAFORSEO_API_KEY new-key-value
/deploy
# Remove a variable
/remove-env-var GOOGLE_ANALYTICS_API_KEY
/deploy- Never commit secrets to git -
.atxp/.env.productionis automatically added to.gitignore - File permissions - Files are created with
600permissions (owner read/write only) to prevent access by other system users - Shell history exposure - Be aware that secrets passed via command line are stored in shell history. For highly sensitive values:
- Manually edit
.atxp/.env.productioninstead - Use environment variables:
/configure-env-var API_KEY "$SECRET_FROM_ENV" - Clear history after use:
history -d $(history 1)
- Manually edit
- Use environment variables for all sensitive data - Don't hardcode credentials in code
- Rotate keys regularly - Use
/configure-env-varto update keys - Audit configured variables - Use
/list-env-varsto review (values are masked)
You can also manually edit .atxp/.env.production:
# Edit the file directly
nano .atxp/.env.production
# Format: KEY=VALUE (one per line)
DATABASE_URL=postgresql://user:pass@host:5432/db
API_KEY=sk-1234567890
ENVIRONMENT=productionAfter manual edits, run /deploy to apply changes.
Deploy agents that require external API access without managing infrastructure:
# Example: SEO analysis agent
/configure-env-var GOOGLE_ANALYTICS_API_KEY ya29...
/configure-env-var DATAFORSEO_API_KEY 1234...
/deployUsers can interact with your deployed agent and pay only for usage.
Deploy agents for your team with shared configuration:
# Configure shared resources
/configure-env-var TEAM_DATABASE_URL postgresql://...
/configure-env-var SHARED_API_KEY sk-...
/deployMaintain different configurations for different instances:
# Staging
/configure-env-var ENVIRONMENT staging
/configure-env-var API_KEY staging-key
/deploy # Creates staging instance
# Production
/configure-env-var ENVIRONMENT production
/configure-env-var API_KEY prod-key
/deploy # Updates or creates production instanceYour instance ID is stored in .atxp-instance after the first deployment:
cat .atxp-instance
# Output: abc123def456This file is used to update existing deployments. Don't commit it to git (it's in .gitignore).
Simply run /deploy again:
# Make changes to your agent
vim my-agent.js
# Deploy updates
/deployThe existing instance will be updated using the ID from .atxp-instance.
To create a new instance instead of updating:
# Remove the instance tracking file
rm .atxp-instance
# Deploy creates a new instance
/deployRun /setup with your connection string:
/setup <your-connection-string>Your .atxp-instance file may reference a deleted instance:
rm .atxp-instance
/deploy # Creates new instanceThe instance ID in .atxp-instance belongs to another account. Remove it:
rm .atxp-instance
/deploy-
Verify they're configured:
/list-env-vars
-
Check the file exists:
cat .atxp/.env.production
-
Redeploy to apply changes:
/deploy
- cloud.atxp.ai - Cloud deployment platform
- ATXP Dashboard - Get your connection string
- ATXP Documentation - Full ATXP docs
For issues or questions:
- ATXP Documentation: https://docs.atxp.ai
- ATXP Support: https://atxp.ai/support
MIT License - see LICENSE for details.