Deploy your first app on Resilis

Sign in to a Resilis account with access to the organization you intend to use. Check your plan and resource limits before creating billable resources. Connecting to the MCP server does not require a paid plan; running resources remains subject to billing and permissions. Keep the agent connection and resource provisioning as separate decisions.

Identify the application root, dependency files, start command, listening port, and environment variables. Decide which data must survive a new application process. The recorded Node.js sample uses memory storage, so restarting it restores its starter tasks. Choose the PostgreSQL sample when you need to follow a persistence check.

• Node.js deployment example →
• Persistent PostgreSQL example →
• Plan prices and limits →

Install Git and use a compatible agent that can read local files and run commands in your workspace. Add the remote Resilis MCP server, sign in through OAuth, and review the requested organization access. Follow the client-specific setup instructions before asking the agent to inspect your app. A configured server alone does not establish a successful deployment.

Ask the agent to inspect tracked, untracked, and ignored files and propose the exact source to upload. Review sensitive files before preparing a commit. The workspace flow uses private Resilis-hosted Git and verifies the selected commit before deployment. You do not need a GitHub repository for this route. Named build contexts are not supported in workspace deployment.

Choose the project, environment, and compute target with the agent. Review build and start settings, required variables, and data connections. Authorize the specific deployment after reviewing that plan. Do not paste credentials into a prompt or commit them to source. The agent must have your authorization for the actions it requests.

• Codex MCP setup →
• Claude Code MCP setup →
• Understand MCP permissions →

In the dashboard, connect the GitHub repository that contains your application. Select the intended branch and app root. Confirm that your connected GitHub account can access that repository. For a sample deployment, fork the linked sample first so that your configuration and changes are under your control.

Review the selected build strategy. Resilis uses Railpack for automatic builds; a Dockerfile route is also available. The Node.js Railpack walkthrough explicitly selects Railpack because its repository also contains a Dockerfile. For the Dockerfile walkthrough, leave command overrides empty so that the image uses the instructions recorded in the guide.

• Deploy from a GitHub repository →
• Deploy a Docker app →

Create a project and environment, then choose a compatible compute target for the app and its database. The recorded PostgreSQL example places both in the same project, environment, and region and uses the internal database connection. Do not assume a private database address is reachable from a different target.

Bind the database connection as a sensitive environment variable, normally DATABASE_URL. Review required values individually; importing detected variable names does not supply their values. Keep secrets out of source, logs, and screenshots. If the app needs object storage, choose managed storage or connect an existing S3, R2, or B2 bucket.

• Configure managed data →

Start the reviewed deployment and inspect its status. Read build logs if compilation or dependency installation fails. For a supported running container, inspect runtime logs if the process exits or requests fail. Use a bounded time range and check the selected environment before interpreting results.

Open the application URL and perform a real action, such as creating and reading a disposable task. Check a readiness endpoint if your app provides one. For the PostgreSQL sample, verify that the same task remains after an app rebuild. Build success and a Running status do not replace an application-level check.

Do I need GitHub? Only the dashboard repository route requires it; the local workspace route uses Git and Resilis-hosted source. Can I attach a custom domain anywhere? No. Custom domains are unavailable on connected compute and S3-backed static sites. Choose a supported managed application target if a custom domain is required.

What happens when I roll back? A supported app rollback selects an application release; it does not restore data or reverse database migrations. What will this cost? Eligible usage consumes plan credit, and usage beyond it is billed separately. Review the billing example before leaving sample resources running.

• Deployment operations and limitations →
• Understand the bill →