Until now, adding AI-powered search to a headless dotCMS application meant working directly with the dotAI REST endpoints — handling authentication, constructing request payloads, and parsing responses by hand. It worked, but it required meaningful setup time on every project. We heard that feedback, and with the latest release of @dotcms/client (v1.3.0), we’ve made it significantly easier.
The new client.ai.search() method gives developers a clean, typed interface to dotAI’s semantic search — no manual REST work required. Pass a natural language query and your index name, and you get back ranked, contextually relevant content results. It supports configurable similarity thresholds, multiple distance functions, content type filtering, and full TypeScript type safety. For React projects, working examples in our Next.js starter show both full-text search and related content patterns you can adapt immediately. It’s available now at no additional charge for all Evergreen headless customers with dotAI enabled.
Learn more in the documentation: