Frequently Asked Questions about AI Suite MCP

AI Suite MCP is a TYPO3 extension that provides a server according to the Model Context Protocol. This allows an AI client like Claude Desktop, Claude.ai, ChatGPT, Claude Code, or Open WebUI to connect directly to your backend and operate there with the same rights as the linked backend user. More about this on the overview page.

The place where work is done. You operate the AI Suite backend module in the TYPO3 backend, while AI Suite MCP brings the same functions into the chat of your AI client. Additionally, there is the structural part that does not exist in the backend module, such as reading page trees, creating, moving, or deleting records, and searching for content.

Yes. AI Suite MCP is a standalone extension but relies on the AI Suite and is installed together with it. The feature and model rights, as well as the connection to external AI providers, come from the AI Suite. Without the AI Suite, AI Suite MCP will not function.

Especially when work is distributed across many pages or records. Typical cases include translating entire page trees, adding missing metadata, recurring text corrections across many elements, and creating page structures. For a single text correction, the backend is usually faster.

Not by default. The preset write mode workspace saves every change in a draft workspace that must be published before it appears on the website. If someone wants it differently, they can switch to auto or live in the configuration. There are two exceptions. uploadMedia and generateImage create physical files and write live in every mode.

TYPO3 from 12.4.11 to 14.3.x and PHP 8.2 or newer. Additionally, the AI Suite in the version that matches the TYPO3 major version, i.e., 12.21.0, 13.15.0, or 14.3.0 and newer. TYPO3 workspaces are installed as a mandatory dependency because the default write mode is based on them.

For the AI-supported tools, yes; for the rest, no. Generation, translation, image creation, and batch processes run through the AI Suite and require a valid key with the appropriate license package. Reading, searching, creating and editing records, text replacements, translations, and embedding media run solely through TYPO3 and work without a key.

Only the tools that call an external AI model consume credits, namely the group of AI functions and the batch processes. All reading tools, all record editing, localizeRecord, and uploadMedia cost nothing. The complete breakdown can be found under available tools.

For batch metadata over a page tree, each page costs credits, and a subtree can quickly be larger than expected. Therefore, a rootPageId is limited to 50 pages. If it exceeds this, the call is rejected before billing, and the found page count is reported. An explicit list of page UIDs remains unlimited because it is compiled intentionally.

The tool remains visible in the tool list, the model calls it, and the call ends with a license error. This is the difference from a missing feature right, as a missing right removes the tool from the list beforehand. If a tool remains visible and fails only upon calling, check the license and not the backend group.

Billing runs through the AI Suite and is independent of whether a request comes from the backend module or an MCP client. Therefore, there is no separate quota for AI Suite MCP. For packages and prices, see packages/licenses.

The endpoint requires productive HTTPS, authenticated via OAuth 2.1 with PKCE or via a bearer token, limits each request to 1 MB, and allows 100 requests per token per minute. With each request, it is checked whether the backend user is still active, and a password change revokes all their tokens. Details can be found under security & operation.

No. A client only sees tools for which the linked backend user has rights, and normal TYPO3 checks on pages, tables, and fields apply during the call. An editor with access to two page trees can edit exactly those two page trees via AI Suite MCP. If someone wants to exclude certain tables fundamentally, even for administrators, they use mcpExcludedTables.

For the tools that call an external AI model, the content to be processed goes to the provider configured in the AI Suite for the respective function. The same providers, models, and data paths apply as in the backend module, see models & APIs. Additionally, the client itself. A chat service running on foreign infrastructure sees everything the model reads through the tools.

AI Suite MCP writes two of its own files. The complete trace with OAuth events, client IP, user UID, and the first approximately 300 characters of each request body, which regularly contains prompts and content excerpts, lands in var/log/aisuite_mcp.log. Only warnings are recorded in var/log/aisuite_mcp_warnings.log. Tokens, hashes, and email addresses are masked before writing. The detailed log can be turned off via mcpLogVerbose. You set the retention period and deletion yourself.

AI Suite MCP processes the same data as the AI Suite and does not add any new providers, so the data protection assessment of the AI Suite remains unchanged. Two points are added. The AI client used is a separate recipient and must be considered, and the detailed MCP log contains personal data and requires a defined retention period.

In the AI Suite backend module in the tab MCP via Revoke Token. This takes effect immediately and regardless of whether the client is still connected. Alternatively, you can remove the connector on the client side. A password change of the backend user also revokes all their tokens.

Yes. The TYPO3 reports module contains the entry AI Suite MCP Security under System → Reports and reports typical misconfigurations, such as allowed HTTP or empty allowlists in the production context.

Guides are available for Claude Desktop, Claude.ai, ChatGPT, Claude Code, the MCP Inspector, and Open WebUI. In principle, any client that speaks Streamable HTTP according to the MCP standard works. An overview with login procedures and accessibility can be found under connecting clients.

That depends on the client. Claude.ai, ChatGPT, and Open WebUI run on foreign infrastructure and call your server from the outside, so they require public HTTPS with a certificate from a recognized CA. Claude Desktop, Claude Code, and the MCP Inspector run locally and can also reach internal hosts, *.ddev.site, and self-signed certificates.

OAuth is the standard way and the only one for externally hosted clients. A static token is set up faster and is suitable for local clients, CI, and scripts, but it expires after mcpTokenLifetimeDays and must then be renewed. OAuth tokens renew themselves.

Yes. Each token is tied to exactly one backend user, and multiple users can connect their own clients simultaneously. Each user only sees their own rights and page trees. A user can also have multiple clients connected in parallel.

Yes, with a local client. For DDEV, the CA must be in the operating system's certificate store, which mkcert -install takes care of. If TYPO3 runs on the same machine as the client, it can also work completely without HTTP, tokens, and OAuth via the local stdio transport.

In most cases, yes. Custom connectors are not automatically active in every conversation. With Claude.ai, ChatGPT, and Open WebUI, you need to enable the connector under the input field per chat; otherwise, the model will not see the tools and will claim it has no access.

The tool list depends on the AI Suite feature rights of the backend group. If a right is missing, the corresponding tool disappears from the list without an error message. If the list is completely empty, it usually means enable_mcp_access is missing. Which right unlocks which tool is listed under permissions & scopes. After adding rights, the client must be re-authenticated.

This is the normal case in the default write mode workspace. Changes are stored as drafts in a draft workspace and only appear after publishing. In the backend, you can see them as soon as you switch to the same workspace. With compareWithLive, a draft can be checked field by field against the live state.

Yes. The command ai-suite-mcp:cleanup should run at least hourly in production, via scheduler or cron. Without it, the table of authorization codes grows indefinitely, session and task directories fill up, and expired tokens remain. See CLI commands.

First in var/log/aisuite_mcp.log. If nothing is there, the request never reached AI Suite MCP, and the access log of the web server is the right source. A 404 indicates a site prefix in the URL, repeated 401 indicates a rejected Authorization header. The most common cases are collected under connecting clients, client-specific ones on the respective subpage.

If your installation is behind HTTP Basic Auth, this conflicts with AI Suite MCP because both use the same Authorization header. Therefore, the MCP paths and the OAuth detection paths must be excluded from Basic Auth. How this looks in the .htaccess is described under production operation.

Not on every one. The extension is designed for PHP-FPM and mod_php, meaning one process per request. It has not been validated on runtimes with persistent workers like FrankenPHP or RoadRunner. If you have multiple application nodes, you need sticky sessions or a shared file system for the session and task directories.