ContentMonster

AI-native content engine — one house running multiple content pipelines that share a common core (AI generation, media ops, QA, multi-tenant config).

• Ad-creative pipeline — Meta (Instagram/Facebook) ad creative, static + video, from a brand brief.
• Book-publishing pipeline — books for Amazon (KDP). (newest realm; design in progress)

Layout (Turborepo + pnpm)

apps/web        Next.js dashboard
apps/cli        commander CLI (pipeline commands, run via tsx)
packages/core   @contentmonster/core — shared engine: ai · media · config · platform · qa · types
clients/<id>    per-tenant config (Jaca Sugar is the test client)
platform/       universal Meta/platform specs
docs/           charter, engineering standards, references, runbooks, state

Commands

pnpm install           # install the workspace
pnpm dev               # run the web app (next dev)
pnpm build             # turbo build
pnpm check-types       # tsc --noEmit across workspaces
pnpm test              # vitest

pnpm --filter cli cli --help          # the ContentMonster CLI
pnpm --filter cli cli steps           # list the ad-pipeline steps

Start here

• What this is + decisions: docs/core-charter/contentmonster-charter.md
• How we build: docs/tools/gstack/use-gstack.md · docs/engineering-standards.md
• Agent project memory: CLAUDE.md

Requires Node >=20 and pnpm (packageManager pins the version). Secrets live in .env* / credentials/ (gitignored) — never committed.