PERSONAIZER

What PERSONAIZER MCP server does

PERSONAIZER MCP server provides MCP tools for creating and using AI chat personas associated with websites. An agent can start persona creation from a website URL, check whether that creation job is progressing, and then interact with the resulting persona through search and chat operations.

The knowledge workflow supports two kinds of input. Plain text can be added as documents, including material such as FAQs, policies, guides, and articles. Product records can also be added or updated with fields for price, stock, attributes, and variants. This makes the server relevant to both general website knowledge and product-oriented support content.

How it works

A typical workflow begins with create_persona, which builds a persona from a supplied website URL. Because creation is represented as a job, check_persona_status can be used to poll its progress rather than assuming the persona is immediately ready.

After the persona is available, agents can use upload_knowledge_text to add plain-text documents or upload_knowledge_docs to add and update product information. search_knowledge performs semantic search over the persona’s knowledge base, while chat sends a message and returns a reply from the persona.

PERSONAIZER MCP server therefore separates persona creation, knowledge ingestion, retrieval, and conversation into distinct tools. An application can use only the parts it needs—for example, search existing knowledge without adding content, or maintain product information before handling chat requests.

Tools and capabilities

• create_persona: Creates a new persona from a website URL.
• check_persona_status: Reports progress for a persona-creation job.
• search_knowledge: Performs semantic search across a persona’s knowledge base.
• chat: Sends a message to a persona and returns its reply.
• upload_knowledge_text: Adds plain text as a knowledge document.
• upload_knowledge_docs: Adds or updates products with price, stock, attributes, and variants.

The tool set supports a website-based persona workflow without requiring all content to come from the original site. Text documents can supplement the website-derived knowledge, and product data can be maintained separately through structured fields.

For developers evaluating fit, the main distinction is between retrieval and conversation: search_knowledge is intended to locate relevant stored information, while chat is the interaction tool for receiving a persona response. The upload tools provide the content-management path, and the creation/status pair handles the initial website-based setup.