Chronomation: Key Architecture Decisions

Chronomation is the engine that powers The Builder Coil. In its first iteration it's a multi-tenant content platform designed to turn raw work artifacts into structured blog posts and social content. Here are the key architecture decisions I made.

Multi-Tenant from Day One

I designed Chronomation to be multi-tenant from the start. Every table has a tenant_id column. Every query is scoped. This adds complexity upfront but makes scaling to multiple customers straightforward.

CREATE TABLE blog_posts (
  id uuid PRIMARY KEY,
  tenant_id uuid NOT NULL REFERENCES tenants(id),
  channel_id uuid NOT NULL REFERENCES channels(id),
  slug text NOT NULL,
  -- ...
  UNIQUE (tenant_id, channel_id, slug)
);

Database: Neon + Drizzle

Why Neon?

• Serverless scaling: No connection pool management headaches
• Branching: Test migrations on branches before production
• Cost: Pay for what you use

Neon is my database host of choice for Chronomation. Learn more at neon.tech.

Why Drizzle?

• Type safety: Schema types flow through your entire app
• SQL-like: Feels like writing SQL, not fighting an ORM
• Lightweight: Minimal runtime overhead

You can read more in the Drizzle ORM docs.

const posts = await db
  .select()
  .from(blogPosts)
  .where(and(
    eq(blogPosts.tenantId, tenantId),
    eq(blogPosts.status, 'published')
  ))
  .orderBy(desc(blogPosts.publishedAt));

Content Ingestion

Chronomation ingests content from multiple sources:

1. GitHub: Commits, changelogs, releases
2. Email: Via Resend inbound webhooks
3. Manual: Notes entered in the admin UI built with shadcn/ui

Each source creates "feed items" that can later be combined into blog posts.

What's Next

In future posts, I'll cover:

• The AI-assisted content generation pipeline
• Social sharing automation
• The admin UI built with shadcn/ui

Follow along as I build Chronomation in public. 🔧

CREATE TABLE blog_posts (  id uuid PRIMARY KEY,  tenant_id uuid NOT NULL REFERENCES tenants(id),  channel_id uuid NOT NULL REFERENCES channels(id),  slug text NOT NULL,  -- ...  UNIQUE (tenant_id, channel_id, slug));