Marquill

Marquill is an AI LinkedIn workspace. Mark researches a topic, drafts posts, carousels, and polls in the user's voice, and publishes them on schedule.

App	Path	Stack	Deploys to
Product app	apps/web	Next.js 16, Clerk, Tailwind v4	Vercel
API and worker	apps/api	NestJS 11, MongoDB, Redis/BullMQ	Railway
Marketing site	apps/landing	Next.js 16 static export	Vercel

Setup

Requires Bun 1.3+ and Docker.

bun install
docker compose up -d        # MongoDB, Redis, BullMQ dashboard (localhost:8080)

# env files, one per app (never at the root)
cp apps/api/.env.example apps/api/.env   # then fill in
# apps/web/.env and apps/landing/.env.local: keys are listed in each app's AGENTS.md

bun run dev                 # web :3000, landing :3001, api :3500

Everyday commands

bun run check               # typecheck + tests across all apps
bun run check:affected      # the same, only for apps changed relative to main
bun run build               # production builds
bun run --filter @marquill/api test   # any script, for one app

Docs

• AGENTS.md: how the apps fit together and the rules for working in this repo. Each app has its own AGENTS.md.
• docs/monorepo.md: the repository layout, deploy settings, Docker, and how this repo was assembled from three.
• apps/api/docs/api/: the client-facing API reference.