FAQ

Answers to the most common questions about vly.ai — what it builds, what it doesn’t, pricing, exporting, and how it compares.

The questions we hear most often, organized so you can skim. Use the table of contents on the right to jump.

About the product

What does vly generate, exactly?

A complete deployed app. The frontend is React + Vite + Tailwind + shadcn/ui. The backend is Convex — a real-time database with serverless queries, mutations, actions, scheduled functions, file storage, vector search, and full-text search. Auth, deployment, and a public URL are included.

Can I see the code?

Yes. Every file is visible and editable in the in-browser editor. Drop into Code mode to make manual edits at any time.

Can I export the code?

Code export is on the roadmap but not currently shipped. Today, your app runs on vly’s managed infrastructure. You connect a custom domain and own the data.

What languages and frameworks does vly use?

React + TypeScript on the frontend (Vite, Tailwind, shadcn/ui, React Router). Convex (TypeScript) on the backend. End-to-end types — the same Project type Convex returns is the type your component consumes. No DTO mapping, no codegen.

Why Convex and not Postgres / Supabase / Firebase?

Convex’s reactive sync (queries that auto-update on the client when underlying data changes) is the killer feature for the kind of apps vly users build — dashboards, collaboration tools, chats, multiplayer. End-to-end TypeScript and serverless functions remove a lot of boilerplate. Read the Convex deep dive for the full case.

You can still use Supabase, Postgres, MongoDB, and others as integrations — they’re just not the primary store.

Is it just a chatbot wrapped around a code template?

No. vly’s pipeline is plan → typed-codegen → build → preview, with a closed-loop type-checker that retries failed builds before showing them to you. The agent has access to your full codebase context, not just snippets, and uses one of three production-grade coding models (Claude Code, OpenAI Codex, Gemini CLI).

Pricing & plans

What does it cost?

Free tier is enough to build and deploy your first few apps. Paid plans start at $3/month — typically 7× cheaper than comparable AI app builders. See plans & pricing for the full breakdown.

What are credits?

Credits meter your agent runs. A typical small change costs 1–3 credits; a large refactor 5–15. Plans include a monthly allotment; you can buy more or upgrade at any time. See Credits.

Is there a free trial?

The Free plan is the trial — no time limit, no credit card. You get a credit allotment to try real builds.

Can I deploy a free-tier app to a custom domain?

Custom domains are a paid feature. Free-tier apps live at *.vly.app. See plans & pricing.

Do you have an enterprise / team plan?

Yes. Team and enterprise tiers add SSO, SCIM provisioning, audit logs, dedicated support, and contracts. Contact us.

Workflow & limits

How long does a build take?

Small changes (1–2 files): 5–20 seconds. Medium features (new page, new schema table): 30–90 seconds. Major rewrites: 2–5 minutes. Plan mode adds 5–15 seconds upfront but typically saves 2–3 round-trips.

Can multiple people work on the same project?

Yes. Projects support real-time collaboration — multiple editors see each other’s changes live, with cursor presence. Branching and review live in Branching.

Can I roll back a change?

Yes. Every prompt produces a new revision in version history. One click to roll back; one click to compare diffs.

Are there rate limits?

Rate limits apply per-workspace based on plan. The Free plan caps concurrent builds and total monthly credits. Paid plans relax both. See Rate limits for the API equivalents.

Can I use it offline?

No — vly runs in the cloud. The agent needs cloud compute and the build pipeline runs on vly’s infrastructure.

AI agents

Which agent should I use?

Default: Claude Code. Best general-purpose pick. Switch to OpenAI Codex for tight, surgical edits and Gemini CLI for multi-file design and schema work. See Choosing an agent for the full decision tree.

Can I switch agents mid-project?

Yes. Choose a different agent for any individual prompt. The whole-project context is shared between agents.

Are my prompts and code used to train models?

No. vly does not use customer prompts, code, or data to train models. Provider zero-data-retention agreements are enabled by default. See Security overview and Data export.

Integrations

Which integrations does vly support?

50+ first-class integrations across payments (Stripe, LemonSqueezy), auth (Clerk, Auth0, WorkOS), email (Resend, SendGrid), AI (OpenAI, Anthropic, Replicate), analytics (PostHog, Mixpanel), and more. See the full Integrations directory.

Can I add an integration that's not on the list?

Yes. Any third-party REST or GraphQL API can be called from a Convex action. For OAuth-based integrations, follow the custom integration guide.

How are API keys for integrations stored?

Encrypted at rest, scoped per environment, never exposed to client code. Set them via the integrations dashboard or vly secrets set. See Secrets management.

Comparisons

vly vs Lovable

Both ship full-stack output. vly’s backend is Convex (reactive by default); Lovable defaults to Supabase. vly is meaningfully cheaper ($3/mo vs higher). vly offers three agent options instead of one. See Migrating from Lovable.

vly vs v0

v0 ships beautiful frontend-only output. You bring your own backend, auth, database, deploy. vly ships everything wired up. Pick v0 if you only need a UI; vly if you need a working app.

vly vs Bolt.new

Bolt focuses on rapid in-browser prototyping. vly focuses on production-ready output with a managed backend, real-time data, and custom domains. See Migrating from Bolt.

vly vs Replit Agent

Replit is a generalist coding environment + agent. vly is specialized for production React + Convex apps with a documented happy path and integrated deployment. Pick Replit for general-purpose dev; vly for product-shipping speed.

vly vs Cursor / Windsurf

Those are AI coding assistants that you run in your existing editor. vly is an AI app builder that owns the full pipeline from prompt to deploy. They complement each other — many vly users use Cursor for hand-editing locally exported code (when that ships).

Support

How do I get help?

Start with Troubleshooting — covers 90% of common issues. Then check the FAQ above. If you’re still stuck, email [e-mail] or join the Discord.

Is there a community?

Yes — the vly Discord is the main hub. Showcase channel, help channel, and weekly office hours.

Can I file a feature request?

Yes — share it in #feature-requests on Discord, or email [e-mail]. The most-requested items make the roadmap.

How do I report a security issue?

Email [e-mail] directly. Do not file public issues. See Security compliance for our disclosure policy.

Still have questions?

Quickstart

Just try it. Five minutes.

Discord

Live community. Office hours weekly.

Email support

Direct line for technical issues.