For Developers — Stack & Lifecycle

A Totalum project is a real, full-stack Next.js application that you own end to end. It is not a sandbox prototype — it's production code with a backend, a database and hosting, built and maintained by the Totalum AI agent as you describe what you want.

#The stack

Totalum projects are built on a modern, fast, maintainable TypeScript stack:

• TypeScript · Next.js · React · Tailwind CSS
• BetterAuth for authentication
• Stripe for payments
• TotalumSDK for the database and platform integrations
• Any npm package — the agent can install and integrate any library

Not supported: PHP, Python, Ruby, Java, .NET, Go, Rust or other non-JavaScript/TypeScript stacks.

#Deployment & scaling

When you click Publish, Totalum automatically deploys your project to Cloudflare — one isolated worker per project, on Cloudflare's global edge network.

• Infinite scalability — the edge platform scales from zero to millions of requests automatically.
• Best-in-class SEO & performance — because projects are server-rendered Next.js on a global CDN, they are fast and fully crawlable (unlike single-page prototypes).
• Custom domains with automatic SSL — connect any domain in a few clicks.

You never manage servers, containers or pipelines — deployment is one click and fully managed.

#The database

Every project ships with a built-in NoSQL database (based on MongoDB) that auto-scales. You consume it entirely through the Totalum Database SDK — CRUD, powerful filtering, nested relations and aggregations — with no connection strings or ORMs to set up. The AI agent creates and maintains the schema as it builds your app.

#You own all your data

Your data is 100% yours, with no vendor lock-in, and you can export every table to JSON at any time using the SDK:

import { TotalumApiSdk } from 'totalum-api-sdk';

const totalum = new TotalumApiSdk({ apiKey: process.env.TOTALUM_API_KEY });

// List every table in the project, then export each one to JSON
const { data: structure } = await totalum.crud.query('nested_data_structure', {});
const tables = structure.results.map((t) => t.table);

const dump: Record<string, unknown[]> = {};
for (const table of tables) {
  const { data } = await totalum.crud.query(table, { _limit: 5000 });
  dump[table] = data.results;
}

// `dump` now holds every record from every table — write it to a JSON file.
console.log(JSON.stringify(dump, null, 2));

#Project lifecycle

1. Create — a project is provisioned with the Next.js template, a database and a sandbox.
2. Build — you prompt the AI agent; it edits the code, updates the database and shows a live preview.
3. Publish — one click deploys to Cloudflare with a public URL (and optional custom domain).
4. Iterate — keep prompting; each change is versioned and can be recovered.
5. Own & export — download the full source code and export all data whenever you want.

Optionally, connect a GitHub repository for a bidirectional code workflow — see Connect to GitHub.

#Platform limits

Totalum projects have generous, auto-scaling limits — 10 TB of storage per project, 5 million rows per database table and 300 tables per project, among others. See the full list on the Limits page.