How to structure multi-project agency codebases with pnpm workspaces and Turborepo, sharing design systems, TypeScript configs, and utility packages with zero disk waste.
A well-structured pnpm monorepo eliminates duplicate code and synchronizes design tokens across your entire digital ecosystem.
When a digital agency or growing company manages multiple related digital products (e.g., a marketing website, a client portal, a mobile API, and an internal admin dashboard), the default instinct is creating separate git repositories for each project. Initially, this seems clean. But as the business expands, maintenance becomes a nightmare: - Updating the company logo or brand color palette requires opening four separate pull requests across four repositories. - Shared TypeScript interfaces (like customer models and invoice schemas) diverge, causing silent API serialization bugs. - Keeping third-party dependencies (like Next.js or React) synchronized across disparate repositories consumes dozens of wasted engineering hours. A monorepo solves this by unifying all applications and shared libraries into a single version-controlled repository.
While npm and yarn support workspaces, pnpm is universally recognized as the gold standard for monorepos due to its unique content-addressable storage model: 1. Content-Addressable Storage: Dependencies are stored once in a centralized global store on your machine. Projects create hard links to the store. If ten projects use React 19, React is stored on disk exactly once, saving dozens of gigabytes of disk space. 2. Non-Flat node_modules: Unlike npm and yarn, which flatten dependencies and create "phantom dependencies" (where code can import unlisted sub-dependencies), pnpm uses symlinks to create a strict, isolated dependency tree. If a package is not declared in your package.json, your code cannot import it, eliminating phantom dependency bugs.
A robust agency monorepo follows this clean organizational hierarchy: root/ ├── apps/ │ ├── web/ # Next.js customer-facing website │ ├── portal/ # Authenticated client portal │ └── docs/ # Technical documentation ├── packages/ │ ├── ui/ # Shared shadcn/ui and Tailwind design system │ ├── config-typescript/ # Base tsconfig.json extensions │ ├── config-eslint/ # Unified linting configurations │ └── utils/ # Shared validation helpers, currency formatters ├── package.json ├── pnpm-workspace.yaml └── turbo.json In pnpm-workspace.yaml, simply declare: packages: - "apps/*" - "packages/*"
As monorepos grow, running linting, testing, and builds across all packages sequentially becomes slow. Adding Turborepo provides intelligent task orchestration and computation caching: When a developer runs "turbo build", Turborepo calculates a cryptographic hash of each package source code and dependencies. If packages/ui has not changed since the last build, Turborepo replays the cached output instantly (0.1s). With remote caching enabled, builds cached by one developer or the CI pipeline are instantly shared across the entire global team.
From custom Next.js engineering and AI automation to high-performance search optimization, Techsist Labs partners with ambitious businesses worldwide to build solutions that scale revenue.
Clear answers to common questions about this topic.
Authoritative documentation, standards bodies, and research benchmarks cited in this analysis.
Official specifications for pnpm workspace protocol and hard link store architecture.
Official technical documentation on building scalable monorepos with Turborepo.
The rise of autonomous buyer agents: how machine-to-machine commerce, programmatic product feeds, and headless checkout APIs are replacing traditional consumer browsing behavior.
A hands-on review of the native generative AI features in Xero (Just Ask Xero / JAX) and MYOB: bank feed reconciliation accuracy, automated GST coding, and where human bookkeepers remain essential.
A financial decision framework for business executives: calculating total cost of ownership (TCO), break-even timelines, and strategic risks between buying commercial SaaS versus building custom AI pipelines.
