Introduction to the Search MCP Tool [BETA]
Essential knowledge
Intended Audience:
Technical User
Author:
Anita Gu
Changed on:
29 July 2026
Overview
The Search MCP Tool bridges the gap between Fluent's documentation and AI tools, enabling seamless access to accurate, up-to-date information. By improving content ingestion and retrieval, it ensures reliable answers and enhances AI-assisted development workflows.Key points
- The Search MCP Tool connects AI tools to thousands of docs articles, and hundreds of rule, util, and component source examples
- It resolves issues like loss of context and unreliable answers caused by poor content ingestion.
- The tool enhances AI-assisted development by allowing tools to pull structured information directly from documentation.
- Users can interact naturally with the documentation, receiving structured summaries and answers to complex queries.
The Search MCP Server now has a tool that connects Fluent's documentation to AI assistants, search tools, and copilots so they can find and use accurate, up-to-date information when answering questions.Fluent Commerce Docs is the repository for best practices and standards that should be followed when developing with Fluent OMS. While humans can naturally scroll through a long webpage and understand the context, AI systems read data differently. Without a dedicated layer to ingest and retrieve this information, LLM-powered tools cannot consume the documentation effectively. As such, when inputting a large amount of web-first content into an agent, it can cause issues such as:
- Loss of Context: Long-form documentation loses its meaning when it is broken down (or "chunked") poorly for the AI to process.
- Unreliable Answers: Because of these ingestion barriers, AI tools cannot reliably answer questions directly from standard documentation today.
How can the Search MCP Server be used?
With an MCP server in place to allow for better ingestion of content, users will be able to:- Converse naturally with the Docs: Instead of clicking through endless menus and pages, users can ask complex questions in their own words and get structured summaries and answers.
- Better AI assisted coding: For developers, this system will allow code assistants to directly pull docs and even source code for existing implementations, resulting in more streamlined code that adheres to best practices automatically, without requiring additional input from the user.
- Query by specific URLs or sections: AI clients can request a precise webpage URL or a single "chunk" of text along with its surrounding context, keeping the AI focused on the exact problem at hand without hitting token limits.
- Inspect source metadata for verification: Every time the AI retrieves information through the server, it returns source metadata. This allows the user or the developer tool to trace the AI's logic directly back to the live, published source URL.
Examples
On asking a coding agent to develop a new rule or component, it can first check whether a compatible rule already exists, minimizing effort and technical debt.
If no such rule or component exists, it can follow examples from the source code of a similar asset when crafting the new one, to ensure the best compatibility with module workflows.
