Prismic MCP. Query structured content and schema data instantly.

You're running into content structure issues? This server connects your Prismic headless CMS straight to your AI agent. You can query and manage document content and schema definitions without ever having to leave your chat or IDE—no jumping back to the main dashboard, I mean.

This setup lets you do way more than just simple searches; it gives your agent deep access into the repo's guts. It handles everything from querying specific fields with raw predicates to listing every single structural element in your content model.

Getting Data: What You Can Pull Down

You can grab the whole payload for any single document using its unique ID via get_document_by_id. This gives you all the component architecture and metadata, wrapped up in one solid JSON blob. If you need to perform a specific search, you'll use query_prismic_documents which executes powerful searches across your entire repo using raw Prismic predicates; this tool handles pagination and lets you filter by locale right out of the box.

Need something more targeted? You can run an advanced search restricted only to a specific language locale—say, Spanish in Mexico—using search_filtered_locale. Or, if you know what type or tag you're looking for, you don't need complex queries. You can list all documents just by their custom content type using list_documents_by_type, or find every document assigned a specific global tag with list_documents_by_tag.

Mapping the Structure: Understanding Your Repo

Before you write any code, you gotta know what's available. This server gives you visibility into the entire content model. You can run get_repo_metadata to pull high-level info about your Prismic repository—it spits out every defined type and every language locale it supports. Want to see everything in one go? Use list_custom_types to get a complete list of all configured custom content types.

Similarly, you can run list_global_tags to pull every single global tag used across the entire system.

It also lists every available language locale (like 'en-us' or 'fr-ca') through list_i18n_languages. If you want to understand how data is filtered, check out get_query_form_schema which fetches the schema details for any specified query form, like checking what fields are involved in an 'everything' search.

The Bottom Line

In short, your AI client uses these tools to perform a full audit of your CMS structure. It reads all custom types and global tags; it checks out the underlying schema definitions for any given query form; it pulls complete document JSONs by ID; and finally, it runs complex, filtered searches using raw predicates against specific types or tags.

You're getting deep structural access to Prismic without needing admin credentials or manual dashboard navigation.