Getting Started

This guide takes you from a new dFlow Cloud account to a running Service. You will create an Application, purchase (or select) a dFlow Worker Node for compute, and deploy.

dFlow is a developer-friendly deployment platform: you keep control of infrastructure (BYOC), and you can later connect an AI agent for deploys, logs, and metrics — see dFlow for Agents when you are ready. Agent setup is not required for this first deploy.

PrerequisitesCopy Anchor Link

• A dFlow Cloud account at app.dflow.sh
• A payment method on file if you will purchase a dFlow Worker Node (plans may include free-node allowances — follow in-product limits)

1. Sign in and open Create New ApplicationCopy Anchor Link

1. Sign in to app.dflow.sh.
2. From your organisation dashboard, choose Create New Application (or Create Application).

2. Application detailsCopy Anchor Link

The wizard opens on Application (Application Details).

1. Enter Application Name (unique in the organisation; follow in-product validation).
2. Optionally add a Description.
3. Continue.

dFlow creates the Application as a draft until the flow finishes. A default production Environment is created for you during the compute step — you do not pick an Environment name in this wizard.

3. Compute — purchase or select a dFlow Worker NodeCopy Anchor Link

The next step is Compute. Workloads run on a Worker Node attached to your Environment.

On dFlow Cloud, open the dFlow compute option and either:

1. Purchase a new dFlow Worker Node — choose a plan (CPU / memory / region), confirm payment, and wait until provisioning finishes; or
2. Select an existing dFlow Worker Node you already own.

You can also attach other compute types later (cloud providers or servers you operate). For this first run, prefer purchase or select a dFlow Worker Node.

Expected outcome: The Environment shows compute attached and ready (or provisioning completes successfully).

If purchase fails (payment method, plan limits), fix the error in-product, then retry. See Billing and Account and Worker Nodes and Compute for deeper guides.

4. Add Services and deployCopy Anchor Link

After compute is ready, the wizard moves to Services (you can skip and add Services later from the Application).

1. Add at least one Service — for example an app from Git, a Docker image, or a database type.
2. Configure source, variables, and any volumes the UI requests.
3. Continue / deploy so the Service builds and starts on your Worker Node.

Expected outcome: A Deployment succeeds and the Service shows as running (or reachable) in the UI.

5. Confirm and open the ApplicationCopy Anchor Link

1. Finish any review / finalize step the wizard shows.
2. Open the Application overview — you should see the production Environment, attached compute, and your Service(s).
3. Open Deployments or Logs on the Service to verify output.

If something failsCopy Anchor Link

• Payment or plan limits on Worker Node purchase → Billing and Account
• Compute stuck provisioning → Compute and worker node issues under Troubleshooting
• Build / deploy errors → Deployment issues under Troubleshooting

Next stepsCopy Anchor Link

• dFlow for Agents — MCP, skills, and plugin overview
• Architecture — Application → Environment → Service model
• Guides — deploy an app, Docker image, database, or template
• Next steps after setup