MCP (Model Context Protocol)

Supported Tools

Available MCP tools for interacting with Unli workspaces

Overview

The Unli MCP server provides tools that enable AI assistants to interact with your workspaces and documents. These tools allow for semantic search, document retrieval, and intelligent information gathering across your connected data sources.

Currently, the following tool is available through the Unli MCP integration:

`search-workspace`

Performs semantic search across all documents in a specified workspace and returns the most relevant results. This tool leverages Unli's vector search capabilities to find contextually relevant information based on natural language queries.

query

string required

The search query in natural language. This can be a question, keywords, or a description of the information you're looking for.

workspace

string required

The identifier of the workspace to search in. This determines which collection of documents will be searched.

top_k

integer

The number of results to return. Default varies by configuration. Use higher values (e.g., 10-20) for broader exploration, lower values (e.g., 3-5) for focused results.

document_internal_ids

array

Optional array of document internal IDs to limit the search scope. When provided, only these specific documents will be searched.

document_name

string

Optional document name to filter results. Searches will be limited to documents matching this name.

Here are example prompts that would trigger the search-workspace tool in different scenarios:

"Search my documentation workspace for information about API authentication methods"

"Find the latest pricing updates in the sales workspace"

"Look up meeting notes from last quarter about the product roadmap"

"What does our employee handbook say about remote work policies?"

"Search for code examples related to database migrations in the engineering workspace"

"Find all references to the new feature launch in project documents"

These prompts demonstrate how to naturally request information, and the AI assistant will use the search-workspace tool to retrieve relevant results from your Unli workspace.

Connect to MCP

Connect your AI tools to Unli AI using the Model Context Protocol

Security Best Practices

Security guidelines and best practices for using Unli MCP