← Back to Glossary

Turborepo

Build Tool

Turborepo is a high-performance build system for JavaScript and TypeScript monorepos. Created by Jared Palmer in 2021 and acquired by Vercel later that year, Turborepo solves a fundamental scaling problem: as codebases grow and split into multiple packages, libraries, and applications within a single repository, build and test times explode. Turborepo uses intelligent caching, task parallelization, and dependency graph analysis to ensure you never rebuild something that has not changed. In a monorepo with a web app, a mobile app, a shared component library, and a utilities package, Turborepo figures out exactly what needs to rebuild when you change a file, runs those tasks in parallel, and skips everything else. For projects with multiple interconnected packages, it can cut CI/CD times by 80% or more.

---

How Turborepo Changed Everything

Before Turborepo, managing JavaScript monorepos was painful. Tools like Lerna (released in 2015 by Sebastian McKenzie, who also created Babel) handled package publishing but offered limited build orchestration. Nx (by Nrwl) provided sophisticated build caching but came with a steep learning curve and a heavy configuration overhead. Many teams simply gave up on monorepos and maintained separate repositories for each package, accepting the trade-off of duplicated configuration, version drift, and cross-repo coordination overhead. Turborepo changed the equation by being radically simple. A single turbo.json configuration file defines your task pipeline, which tasks depend on which other tasks, which tasks can run in parallel, and what outputs to cache. Run turbo build once, and every subsequent build that has not been invalidated by a file change completes in milliseconds because Turborepo replays the cached output. Jared Palmer's insight was that most developers do not need a complex build orchestrator; they need a fast, smart cache that understands their dependency graph. The Vercel acquisition in December 2021 gave Turborepo access to remote caching infrastructure, allowing teams to share build caches across machines and CI runners.

One Thing Most People Miss

Turborepo's remote caching is the feature that separates it from a local optimization tool and makes it a team-level productivity multiplier. When one developer runs a build, the output artifacts get uploaded to a shared remote cache. When another developer, or a CI runner, runs the same build with the same inputs, Turborepo downloads the cached output instead of rebuilding from scratch. This means the first developer to build on Monday morning effectively pre-warms the cache for the entire team. In practice, this means new developers cloning a large monorepo can have all packages built in seconds instead of waiting ten minutes for a cold build. CI pipelines that used to take fifteen minutes complete in under two. The cache is content-addressed, meaning it is based on file hashes, not timestamps, so it is deterministic and reliable. For client projects that grow into multi-package monorepos, a web app, an admin dashboard, shared UI components, and a shared API client, Turborepo keeps the development experience snappy regardless of how large the codebase gets.

Visit: turbo.build