MCP Ask SISTRIX

Embedded in: https://www.sistrix.com/ask-sistrix/ai-basics/mcp/
Which AI clients are supported?

The SISTRIX MCP server works with Claude and ChatGPT. Gemini and Microsoft Copilot are not supported as of yet. As soon as these platforms offer the integration of MCP servers, SISTRIX will be ready. For combined packages, like Langdock, the availability of all tools is dependent on the corresponding configuration […]

Which SISTRIX data can the MCP server access, and which is excluded?

The MCP server has access to keyword rankings, Visibility Index, keyword lists, competitor data and the GEO & Visibility Tools (prompt research, AI check, prompt tracking). It cannot access Google Search Console data, even if you set up a GSC connection within SISTRIX. Since SISTRIX only hands over the GSC […]

How high is the quota and what does the MCP usage cost?

From our side, there is no separate MCP limit. The connection is included in the SISTRIX package without further retrieval costs. Rather, the costs originate from the AI platform: Claude and ChatGPT run on their own subscriptions or API usage. The limit that you may reach with intense use thus […]

Can I conduct recurring retrievals or dashboards via the MCP server?

Yes, for instance as HTML output. Since dashboards need fresh data constantly, retrievals have to be performed anew regularly. Caching is not effective here. For very frequent queries, such as daily ones for big domains, you may reach the limit of the AI platform quickly.

Will the MCP run automatically or do I have to explicitly mention it in the prompt?

This depends on the client. In Clause Desktop, the MCP server will be automatically activated as soon as a connection is established. Claude decides itself when a retrieval is sensible. However, you can also explicitly (de)select access yourself. A proven approach: create one project per use case in Claude, add […]

Which model is recommended and how do I get better outputs?

For most analyses, Claude Sonnet offers a good price/efficiency ratio. Opus is more powerful but for many use cases not necessary. The quality of outputs is more affected by prompts, rather than the choice of model: concrete instructions regarding the output format, like tables, structured lists or HTML will lead […]

How can I retrieve keywords, search volume, and clusters with the MCP?

Keyword data, search volume, cluster analyses including ranking URLs and questions per keyword can be all retrieved via the MCP server. For very long keyword lists, errors may occur since the server scans keywords individually. Recommendation: separate the list into smaller batches.

What are the GEO & AI Visibility tools, and what can the MCP accomplish in connection to them?

Prompt Research, AI Check, and Prompt Tracking are included in all SISTRIX packages and can be fully utilised via the MCP server. You can find a good introduction for GEO analyses with MCP here: https://www.sistrix.com/ask-sistrix/ai-basics/mcp/geo-analysis/.

Can the MCP be combined with other data sources?

Yes. In one prompt you can combine data from multiple MCP servers, for example SISTRIX data with a Screaming Frog crawl. Additionally, the MCP can be incorporated well with AI skills and agents, which enables a much better workflow than the pure usage of prompts.

Where can I find use cases and prompt templates?

We summarised the different use cases and compiled them for you here: Initial SEO analysisContent Gap AnalysisContent AuditGEO AnalysisAutomatic Keyword Strategy