Quickstart

You'll need a GitHub account with a repository to deploy, a free Vercel account, and a free Render account. deplo.ai deploys into accounts you own — it never hosts your app itself.

Deploy your first app#

1. Sign in with GitHub

   Go to deplo.in and click Sign in with GitHub. Signing in creates your account — there is no separate registration step. deplo.ai requests repository access so it can read your repo's file tree during analysis and create deploy webhooks.

2. Connect Vercel

   After sign-in you land on onboarding, which asks you to authorize the deplo.ai Vercel integration. This is a one-click OAuth flow — deplo.ai receives a token scoped to creating and deploying projects in your Vercel account, and stores it encrypted. Your frontend will appear in your Vercel dashboard like any other project.

3. Connect Render

   Render has no third-party OAuth, so you connect it with a Personal API key: open Render → Account Settings → API Keys, create a key (it starts with rnd_), and paste it in Dashboard → Integrations. The key is validated live against the Render API before being stored encrypted — an invalid key is never saved.

4. Pick a repository

   Open Dashboard → Repositories. Your GitHub repos sync automatically; hit Sync if a fresh one is missing. Private repositories additionally require installing the deplo GitHub App — the dashboard prompts you if it's needed.

5. Analyze

   Click Deploy on a repository. deplo.ai reads the repo tree and detects the stack — frameworks, monorepo layout, build commands, runtimes — using a deterministic detection engine, with an AI fallback for unusual layouts. You'll see the detected frontend/backend and where each lives before anything deploys.

6. Provide environment variables

   The analyzer scans your source for environment variables the code actually reads — not just .env.example — and marks which are required. Fill in the values; connection-wiring variables like NEXT_PUBLIC_API_URL and FRONTEND_URL are injected automatically, so you never provide URLs that don't exist yet.

7. Deploy

   Confirm, and watch the live log. The backend deploys first to your Render workspace; its live URL is then wired into the frontend build, which deploys to your Vercel account. A typical full-stack deploy completes in two to four minutes.

   deployment logcopy

   Analyzing repository…
   Detected: frontend=Next.js (client/) · backend=Express (server/)
   Deploying backend to Render…
   Backend live: https://deplo-my-app-api.onrender.com
   Injecting NEXT_PUBLIC_API_URL into frontend build
   Deploying frontend to Vercel…
   Frontend live: https://deplo-my-app-web.vercel.app
   Deployment completed

8. Verify and monitor

   Open the frontend URL — your app is live. From the deployment page, Project monitoring & analytics gives you uptime, latency, and server CPU/memory for the new deployment, plus optional privacy-friendly product analytics via a one-line snippet. See Monitoring & Analytics.

If the deployment fails#

Failures are treated as a first-class experience, not a dead end:

• The deployment page shows the full log plus an AI diagnosis: the likely cause, the affected file, and a fix.
• You get an email with the same diagnosis and a copy-paste prompt for your AI coding tool (Cursor, Claude Code, Copilot) that describes the failure precisely.
• Fix, push, and hit Retry — the platform reuses the same providers' services idempotently.

Redeploying#

After the first successful deployment, pushing to the deployed branch triggers a redeploy automatically — a webhook is registered on your repository, and the previous configuration (detected stack, env vars) is reused so pushes go straight to deploy. You can also redeploy any time with one click in the dashboard or one deplo command; existing Vercel projects and Render services are reused idempotently.

Next steps#

• Install the CLI and deploy without opening the dashboard.
• Understand the pipeline — what deplo.ai does at each stage.
• Master environment variables — scopes, detection, wiring.