Skip to main content

​
The Deploy Command

Once you’ve built and tested your agent locally, deploy it to Blink Cloud with a single command: 
blink deploy
This command:
  1. Builds your agent for production
  2. Uploads your compiled code and source files to Blink Cloud
  3. Syncs environment variables from .env.production
  4. Creates a new deployment version
  5. Makes your agent live and accessible to users

Blink Cloud Features

Learn about logging, tracing, debugging, and monitoring features available in Blink Cloud.

​
First-Time Deployment

Deployment When you run blink deploy for the first time, Blink will:
  1. Prompt for organization: Select which organization should own this agent (or automatically use your only organization)
  2. Create the agent: Blink creates a new agent using your package.json name
  3. Save configuration: A data/config.json file is created to remember your deployment settings
The data/config.json file contains no secrets and can be committed to source control. It links your local agent to your cloud deployment.

​
Configuring Environment Variables

Environment variables are how your agent accesses API keys, secrets, and configuration in production.

​
Local Development vs Production

Blink uses different environment files for different contexts:
  • .env.local - Used during blink dev for local development only
  • .env.production - Deployed to Blink Cloud with your agent
1

Create .env.production

Create a .env.production file in your project root with the variables your agent needs in production:
ANTHROPIC_API_KEY=sk-ant-...
OPENAI_API_KEY=sk-...
GITHUB_TOKEN=ghp_...
2

Deploy your agent

Run blink deploy. Blink automatically uploads all variables from .env.production to Blink Cloud.
blink deploy
3

Verify deployment

Blink compares your .env.local and .env.production files. If any variables exist in .env.local but not in production, you’ll see a warning:
Warning: The following environment variables are set in .env.local
but not in .env.production:
- SLACK_BOT_TOKEN
- DATABASE_URL

Do you want to deploy anyway?
This helps catch missing configuration before your agent fails in production.
Environment files are ignored by default. When you create a new agent, .env files are automatically added to .gitignore to keep your secrets safe.

​
Webhooks in Production

If your agent handles webhooks (from Slack, GitHub, etc.), Blink provides a production webhook URL after deployment.

​
Local Development Webhooks

During blink dev, Blink provides a temporary webhook URL via devhook: 
blink dev
# Output: ⚙ Send webhooks from anywhere: https://abc123.blink.host
This URL forwards webhooks to your local machine for testing.

​
Production Webhooks

After running blink deploy, your agent gets a permanent webhook URL.

​
Automatic Webhook Migration

On your first deployment, if you’ve been using devhook during local development, Blink will prompt you to migrate your webhook tunnel to production: 
blink deploy

Webhook Tunnel
  Current: https://abc123.blink.host → local dev
  After: https://abc123.blink.host → production
  Migrating will keep your webhooks working in production

? Migrate tunnel to production? (Y/n)
If you choose Yes:
  • Your devhook URL (https://abc123.blink.host) automatically points to your production agent
  • Your existing webhook integrations (Slack, GitHub, etc.) continue working without reconfiguration
  • You don’t need to update webhook URLs in external services
If you choose No:
  • Your agent gets a new production webhook URL
  • You’ll need to manually update webhook configurations in external services
To continue developing locally with webhooks after migration, you’ll need to reconfigure external services to point back to your new devhook URL when running blink dev.

​
Deployment Workflow

Here’s the recommended workflow for deploying updates:
1

Develop locally

Build and test your changes using blink dev. Use the temporary devhook URL for testing webhooks.
2

Update .env.production

If you added new environment variables to .env.local, add them to .env.production as well.
3

Deploy

Run blink deploy to create a new deployment. Optionally, add a message describing the changes:
blink deploy --message "Add GitHub issue creation tool"
4

Monitor deployment

Blink polls until your deployment is live. If it succeeds, all users immediately start using the new version.If deployment fails, Blink shows the error and provides a link to view logs in the dashboard.
5

Verify in production

Test your agent in Blink Cloud to ensure everything works as expected. Check the deployment logs if you encounter issues.

​
Deployment Versions

Each blink deploy creates a new numbered deployment (e.g., #1, #2, #3). This allows you to:
  • Track changes: See what changed in each deployment
  • Roll back: Revert to a previous version if issues arise
  • Compare performance: Measure improvements across versions
View all deployments in the Blink Cloud dashboard under your agent’s Deployments tab.

​
Troubleshooting

​
Deployment fails with “Missing environment variable”

Your agent code references an environment variable that isn’t set in .env.production or the Blink Cloud dashboard. Add the missing variable and redeploy.

​
Webhooks not working in production

Verify that:
  1. You’ve updated the webhook URL in the service’s settings (Slack, GitHub, etc.)
  2. Your agent’s request event handler is properly configured
  3. The webhook URL is correct (check the output after blink deploy)

​
Changes not appearing after deployment

Check the deployment status in the Blink Cloud dashboard. If the deployment succeeded but you don’t see changes:
  • Clear your browser cache
  • Verify you’re testing in the correct organization/agent
  • Check if you’re in a different chat that might have cached state
⌘I