AIData Systems & AI LabProduct Analytics · Data Systems · AI Workflows · Decision InfrastructureContact
Open OS launcher

Deployment / operations

ENV / Config

ENV and config management separates public app settings from private secrets so AI, auth, database, collaboration, and deployment services can run safely.

systems.env

What this system does

Separates public client configuration from private server-only keys.

Keeps local, preview, and production deployments consistent without hardcoding credentials.

Makes debugging easier by grouping auth, database, AI provider, collaboration, blob storage, and workflow variables.

How I use it

Ghost-AI uses environment-backed services for Clerk auth, Postgres/Prisma, Liveblocks, Trigger.dev, OpenRouter, and Vercel Blob.

The portfolio documents configuration patterns with masked values only, not real keys.

Deployment thinking includes Vercel build context, production debugging, and making routes safe when credentials are missing or private.

examples.evidence

Evidence or examples

Ghost-AI package includes Clerk, Liveblocks, OpenRouter, Prisma/Postgres, Trigger.dev, and Vercel Blob dependencies.

Ghost-AI source references DATABASE_URL, LIVEBLOCKS_SECRET_KEY, NEXT_PUBLIC_CLERK_SIGN_IN_URL, NEXT_PUBLIC_CLERK_SIGN_UP_URL, TRIGGER_PROJECT_REF, OPENROUTER_API_KEY, and BLOB_READ_WRITE_TOKEN.

No secret values were copied, printed, or added to the portfolio.

.env.example - AI and workflow

env
OPENROUTER_API_KEY=********
TRIGGER_PROJECT_REF=<configured>
TRIGGER_SECRET_KEY=********

.env.example - auth and collaboration

env
NEXT_PUBLIC_CLERK_SIGN_IN_URL=/sign-in
NEXT_PUBLIC_CLERK_SIGN_UP_URL=/sign-up
CLERK_SECRET_KEY=********
LIVEBLOCKS_SECRET_KEY=********

.env.example - data and artifacts

env
DATABASE_URL=postgresql://<configured>
BLOB_READ_WRITE_TOKEN=********
NEXT_PUBLIC_SUPABASE_URL=<configured>

Public vs private config rule

pattern
NEXT_PUBLIC_* can reach the browser. Secret keys, database URLs, blob tokens, and provider keys stay server-side only.
Content is evidence-first. If a system detail is conceptual, it is framed as a system focus or implementation pattern rather than a fake production claim.