§ Three Deployment Modes

One platform. Three businesses.

Because the platform collapses the toolchain into a single database-driven runtime, the same core technology can be deployed in three fundamentally different configurations. Each serves a different market. Each unlocks a different category of value. The three reinforce each other.

SaaS for Individual Developers

A multi-tenant SaaS platform where developers build and deploy applications without configuring a single piece of infrastructure.

Who it’s for

Individual developers, freelancers, small teams, hackathon founders. Anyone whose first week of a project usually dies in environment setup. Anyone who has lost a Saturday to a npm install they did not understand.

The promise

Onboarding measured in minutes instead of weeks. Sign up. Open the browser. Build. There is nothing to install, no local environment to configure, no repository to initialize. Platform improvements reach every user simultaneously through database updates — no version fragmentation, no breaking-change migrations.

Key metric

Time to first deploy: under 60 seconds.

Self-Hosted Docker for Enterprises

Self-hosted Docker instance inside the customer’s own infrastructure.

Who it’s for

Regulated enterprises — healthcare, finance, government, defense — and any organization where IP isolation is non-negotiable. HIPAA, SOC 2, FedRAMP, GDPR, CCPA. Data residency requirements. Internal IP that cannot leave the corporate VPC. Customers currently spending millions per year stitching together a dozen vendors and a DevOps team to keep them running.

The promise

Because the platform is defined by a database schema and a browser-based runtime, a siloed instance deploys into AWS, Azure, GCP, or an on-prem data center as a self-contained Docker image. The enterprise gets full control of code, data, and access policies — without sacrificing the productivity benefits of the SaaS configuration.

Key metric

$450K–$700K / yr eliminated per 10-developer team.

Agentic AI Runtime

An environment designed to be as accessible to an AI agent as it is to a person.

Who it’s for

AI labs and orchestration teams building autonomous developer agents — Cursor, Claude Code, Devin-class systems and beyond. Every traditional developer environment was designed for humans and is being retrofitted for AI. This one was designed in a way that happens to work as well for an agent as it does for a person.

The promise

The full development lifecycle — read, write, execute, test, version, deploy — collapses into database transactions. An agent reads code through a query, writes through an insert, executes through hydration, tests by running the deployed component against expected behavior, and iterates by rewriting records when tests fail. No shells. No file systems. No Git. No CI/CD APIs. No Kubernetes manifests. Promotion between environments is one SQL statement away from atomic success or atomic rollback.

Key metrics
  • 1 — Tool surfaces the agent must master (the database)
  • 0 — Failure modes from shell / fs interaction
  • Trivial — Deployment cost per iteration

The Three Form a Spectrum

Individuals build on SaaS. Enterprises consolidate. Agents operate autonomously.

All three rest on the same foundation: code as data, deployment as a transaction, browser as operating system. Continue to the differentiators to see how this compares to what already exists.