How is this different from connecting Prismic MCP directly?

Connected directly, an agent sees every tool the server exposes and holds the credential itself. Through TrustGate you choose which tools each agent gets, and every call is logged against the agent that made it.

What if the Prismic MCP API changes?

We track the Prismic MCP schema, so the tool catalogue stays current. Tools you have already allowed keep working, and anything new stays denied until you enable it.

Can I give different agents different Prismic tools?

Yes. Access is set per agent: one can get a handful of read tools, another the full set. You grant the whole MCP or single tools within it.

Where can I see what my agents did in Prismic?

Every call is logged against the agent that made it, with the tool it used. Activity can be sent on to your own monitoring.

Do I need to run the Prismic MCP server myself?

No. TrustGate connects to the Prismic MCP server and sits in front of it. Your agent only talks to your TrustGate URL, never to Prismic directly.

Connect Prismic MCP to your AI agents, securely.

Q: Is Prismic MCP secure?

The server itself exposes 16 tools, and some of them can write to your Prismic account. Behind TrustGate an agent only sees the tools you allow, requests and responses are inspected, and every call is logged. Anything outside the allowlist is denied before it reaches Prismic.

One secured endpoint for every Prismic call your AI agents make, scoped per user, inspected, and logged.

Get started for free

Get a demo

Try Prismic MCP now

What your agents can ask once Prismic MCP is connected.

“Which custom types and shared slices does our repository define?”
“Find all blog posts tagged product-launch in the French locale”
“Create a release for the spring campaign”

“Draft a new landing page in that release using the hero slice”
“Upload this image to the asset library for the new page”
“Show me every version of the pricing page”

Select the tools your AI agents can call

16 tools from the Prismic MCP schema.

Showing 16 tools

list_repositoriesreadList all Prismic repositories you have access to. Called first to discover available repositories.
list_custom_typesreadList all custom types (content models) in a repository with their id, label, and format.
list_shared_slicesreadList all shared slices in a repository with their variation names.
list_localesreadList all locales configured in a repository.
get_custom_typereadFetch a custom type's full field model and an empty content template for creating documents.
get_shared_slicereadFetch a shared slice's full model with all variations and ready-to-use empty variation bodies.
search_documentsreadSearch documents by keywords with optional filters for custom type, locale, status, and tags. Returns metadata, not full content.
get_documentreadFetch a document's metadata and full content. Supports reading from a specific version or release.
list_document_versionsreadList every version of a document, including drafts, published, archived, and release versions.
create_documentwriteCreate a new document inside a release. Missing fields are filled with model defaults.
update_documentwriteUpdate a document into a release with full content. The new version is branched from a base version.
search_assetsreadSearch the asset library by keyword. A single query matches against filename, alt text, notes, and credits. Returns previews.
upload_assetwriteFetch an asset from a URL and upload it to the repository's asset library. Returns the Prismic CDN url for use in documents.
list_releasesreadList all releases in a repository, newest first.
create_releasewriteCreate a new empty release. Returns its id for use with create_document and update_document.
publish_releasehigh riskPublish a pending release, making all of its documents live. Requires a Publisher (Manager) role. Irreversible.

What Prismic MCP exposes

Prismic MCP groups its 16 tools into 4 categories, each with a suggested starting policy.

Category	Covers	Access	Suggested policy
Repository & Models	Listing the repositories the user can access, custom types, shared slices and locales, and fetching a custom type's full field model or a shared slice's variations with empty templates	Read	Read-only and the prerequisite for valid writes, since create_document fills missing fields from the model. Access follows the signed-in user, so an agent only reaches repositories that account can already open
Documents	Searching documents by keyword, custom type, locale, status and tags, reading a document's full content from a specific version or release, listing its versions, and creating or updating documents inside a release	Read / Write	Writes never touch live content: create_document and update_document only work inside a release, and updates branch from a base version. That makes this group safe to allow for drafting, provided publishing stays gated
Assets	Searching the asset library by filename, alt text, notes and credits, and fetching a file from a URL into the asset library	Read / Write	upload_asset pulls from any URL the agent supplies and returns a public CDN url, so log the source URL and restrict it to agents that are meant to add media
Releases	Listing releases, creating empty releases to stage changes in, and publishing a pending release so all of its documents go live	Read / Write	publish_release is irreversible, cannot be undone through the server and needs a Publisher (Manager) role. Prismic recommends reviewing staged changes in the dashboard instead of publishing through the assistant, so keep it behind approval or deny it