Skip to Content
Getting started

Getting started

Ship your first ZXCV app in four steps: create a project, add a table, query it with the SDK, and deploy to the edge.

AI Assistance
Help me start a new project on ZXCV. Do the following:
1. Install the CLI: run "npm i -g zxcv" then "zxcv login".
2. Create a project and a "todos" table (id, title, is_complete) with a couple of sample rows.
3. Install @zxcv.build/sdk and initialize a client with my project ref + anon key from Settings → API.
4. Fetch and render the todos in my framework.
5. Deploy with "zxcv deploy --prod".
Ask me which framework I'm using before writing any code.

ZXCV is in public beta at zxcv.build  — no card required.

Prerequisites

  • A ZXCV account
  • Node.js 20+ (only for the CLI workflow)

Create a project

From the dashboard, create a new project. ZXCV provisions a Postgres database, an edge project, and a zxcv.app subdomain in a few seconds.

Add a table

Create a table in the SQL editor — for example a todos table with id, title, and is_complete. Copy your project ref and anon key from Settings → API.

Query it from your app

Install the SDK and read your data — typed from your schema, filtered by your policies:

import { createClient } from '@zxcv.build/sdk'
 
const zxcv = createClient({ project: 'my-app', anonKey: process.env.ZXCV_ANON_KEY! })
 
const { data } = await zxcv.from('todos').select('*')

Pick your framework in the quickstarts for the full setup.

Ship it

Deploy to the global edge from the dashboard, or from your terminal:

npm i -g zxcv && zxcv link my-app
zxcv deploy --prod        # → my-app.zxcv.app · edge · 300+ PoPs
zxcv deploy --canary 10   # 10% of traffic, auto-promote on clean metrics

Next steps

  • Auth — the 3-slot model and the .policy language
  • SDK reference — @zxcv.build/sdk, typed from your schema
  • Connect a coding agent with zxcv mcp for full project context
IntroductionNext.js