Meta Unveils Docusaurus 3.9 with AI-Enabled Search

Meta’s open‑source documentation framework, Docusaurus, just dropped its 3.9 release. The update brings AI‑powered search via Algolia DocSearch v4, new runtime optimizations, and expanded internationalization support.

AI Search: A Game Changer

Algolia’s DocSearch v4 now harnesses generative AI to surface context‑aware answers. Instead of returning a list of URLs, the search can summarize relevant sections, making onboarding faster for developers.

Users can toggle between traditional keyword search and AI mode from the UI, ensuring flexibility during migration.

Modernized Runtime and Performance Boost

Version 3.9 removes legacy Node 18 support and adopts ESM‑only build pipelines. The change reduces bundle size by ~15% and aligns Docusaurus with modern CI/CD workflows.

• Improved server‑side rendering via Webpack 5.
• Enhanced client‑side hydration with React 18’s concurrent features.
• Zero‑config incremental builds for large monorepos.

Expanded Internationalization (i18n)

The new i18n API lets sites declare multiple locales in a single config file, supporting nested language trees and fallback chains.

“Multi‑lingual docs used to be a pain; now we can ship them with a single command,” says Meta engineer Maria Lopez.

Multi‑Domain Deployment Made Simple

Organizations can host distinct documentation sites under the same GitHub repo. Docusaurus 3.9 introduces siteConfig.domains for clean domain routing and shared authentication.

1. Define domain patterns in docusaurus.config.js.
2. Deploy via Vercel or Netlify; the runtime auto‑detects the target domain.
3. Maintain a shared navigation bar across sites.

Upgrade Path & Compatibility

Meta recommends a two‑step upgrade: first bump to 3.8 to resolve deprecations, then migrate to 3.9. The new release drops support for Node 18, so projects on that runtime must move to Node 20 or later.

Migration scripts and a detailed changelog are available on GitHub.

What Comes Next?

Meta is already prototyping a plugin that lets teams inject proprietary knowledge bases into Docusaurus’s AI search. The company plans a 4.0 release in early 2026, adding native GraphQL API support for docs content.

Developers can preview the 3.9 beta on GitHub and report issues via the project’s issue tracker.