Frequently Asked Questions

Billing & accounts#

How much does deplo.ai cost?#

deplo.ai is free while in beta. Your apps run on your own Vercel and Render accounts, so the only costs are whatever those providers charge you — and both have free tiers that comfortably run side projects and demos.

Do I need paid Vercel or Render plans?#

No. Backends are created on Render's free plan by default, and frontends deploy to Vercel's Hobby tier. Free Render services sleep after inactivity and wake on the next request — upgrade the service in your own Render dashboard if you need always-on.

How do I delete my account?#

Settings → Danger Zone. Because your apps live in your own provider accounts, deleting your deplo.ai account never takes a deployed site down — it only removes deplo.ai's access and data.

Providers#

Why do deployments go to my accounts instead of deplo.ai hosting them?#

It's the core design decision: no lock-in. You keep the provider relationship, the URLs, the logs, the billing, and the ability to walk away. deplo.ai is the orchestration layer, not the host — if deplo.ai disappeared tomorrow, every deployed app would keep running untouched.

Can I use a custom domain?#

Yes — on the provider's side. The Vercel project deplo.ai creates is a normal project in your account, so you add domains in your Vercel dashboard exactly as you would for any project. Same for Render.

Which providers are supported?#

Vercel for frontends, Render for backends — today. See the provider roadmap for what may come next.

Deployments#

Which frameworks can I deploy?#

Frontends: Next.js, React/Vite, Vue, Nuxt, Astro, Svelte, SvelteKit, Remix, Gatsby, Angular, and static HTML. Backends: Express, Fastify, FastAPI, Flask, Django, and Go. The full matrix with detection details is in the introduction.

How long does a deployment take?#

Typically two to four minutes for a full-stack repo — dominated by the providers' build times, not by deplo.ai. Backend-only or static deploys are faster.

Can I roll back to a previous deployment?#

Not yet from deplo.ai — rollbacks are planned. Today you can roll back in the provider's own dashboard (Vercel keeps every deployment; Render keeps previous deploys), or redeploy an older commit.

Does deplo.ai deploy automatically when I push?#

Yes — after a repository's first successful deployment, a webhook redeploys it on every push to the deployed branch, reusing the detected stack and env vars. For pipelines that need gating (tests before deploy, exit codes), run the CLI in CI instead.

My deployment failed. Where do I start?#

The deployment page shows the full log with an AI diagnosis — cause, affected file, fix — and the failure email includes a prompt you can paste into Cursor, Claude Code, or Copilot. Start at Troubleshooting.

GitHub#

Why can't deplo.ai see my private repository?#

OAuth sign-in covers identity and public repos; private repositories additionally require installing the deplo GitHub App on your account or org. The dashboard prompts you when it's needed — details in the GitHub integration docs.

Does deplo.ai clone my whole repository?#

No. Analysis reads the file tree and a bounded set of small config and source files through the GitHub API. Builds happen on the providers, which pull the repo themselves.

CLI#

Is the CLI different from the dashboard?#

No — it's a thin client over the same API. Anything the CLI deploys shows up in the dashboard and vice versa. Install with npm i -g @deplo.ai/cli; see the CLI docs.

Why is the package @deplo.ai/cli and not deplo?#

The bare deplo name on npm belongs to someone else. Always install @deplo.ai/cli — the installed command is still deplo.

Security#

How are my provider credentials stored?#

Encrypted with AES-256-GCM at rest, validated before storage, never logged, never shown again after you enter them. Deployments run with your credentials against the providers' APIs — deplo.ai holds no global deploy keys. Full details in Authentication & Security.

What can deplo.ai do with my Render API key?#

Create and deploy web services in your workspace, set their environment variables, and read service metrics for monitoring. You can revoke the key in Render at any time; deplo.ai's copy becomes useless the moment you do.

Is the analytics script tracking my users?#

The optional insights snippet is cookieless and anonymous: it records path and referrer with a visitor hash that rotates daily and honors Do Not Track. No cross-site tracking, no personal data. See Monitoring & Analytics.

Architecture & performance#

Is AI making the deployment decisions?#

Mostly no — detection is deterministic (weighted framework signatures over your repo tree) and AI is a bounded fallback: at most two model calls per deployment, used only when confidence is low or a failure needs diagnosing. A wrong AI answer can never break a working deterministic result.

Why is my backend slow on first request?#

That's Render's free-plan cold start, not deplo.ai — free services sleep when idle. Upgrade the service plan in your Render dashboard for always-on.

Where does deplo.ai itself run?#

The same way your apps do: the dashboard on Vercel, the API and workers on Render, Postgres and Redis as managed services. We deploy deplo.ai with deplo.ai where possible — the terminal on the landing page is a real deploy.