> ## Documentation Index
> Fetch the complete documentation index at: https://acem-52171079.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Tools

> Capabilities exposed to AI agents

The MCP Server exposes a set of tools that allow agents to read, write, and query data within the DataDot ecosystem.

## Discovery

### `search_tools`

Dynamically discovers available tools based on a query.

* **Args**: `query` (str), `limit` (int, default: 5)
* **Use Case**: "What tools can I use to manage files?"

## RAG & Retrieval

These tools allow the agent to tap into the knowledge base.

### `query_content`

**The most powerful tool.** Performs a semantic search across all documents in the workspace and returns context-aware answers.

* **Features**: Supports streaming reasoning, source citation, and relevance scoring.
* **Args**: `query` (str), `mode` ("query" or "chat"), `top_n` (int)
* **Use Case**: "What does the architecture document say about authentication?"

### `search_files`

Performs a semantic or keyword search to find relevant files without retrieving their full content.

* **Args**: `query` (str)
* **Use Case**: "Find all files related to 'onboarding'."

### `get_citation`

Retrieves specific citation metadata for a document ID, ensuring that the agent can attribute information correctly.

* **Args**: `doc_id` (str)
* **Use Case**: "Get the publication date for document 'doc-123'."

### `get_embeddings`

Retrieves vector embeddings and metadata for a specific document or all documents in the workspace.

* **Args**: `document_id` (optional str)
* **Use Case**: "Check how many vectors are stored for this file."

## Document Management

Tools for modifying the workspace and file system.

### `process_file`

Triggers the embedding and indexing process for a specific document. This performs text extraction, chunking, and vector storage.

* **Args**: `document_id` (str), `use_cache` (bool)
* **Use Case**: "Process the uploaded PDF."

### `delete_file`

Removes a file from the workspace index and database.

* **Args**: `document_id` (str)
* **Use Case**: "Remove the outdated policy document."

### `get_file_content`

Reads the raw text content of a file.

* **Args**: `document_id` (str), `limit_bytes` (int)
* **Use Case**: "Read the content of 'config.json'."

### `get_source_file`

Lists all files currently available in the workspace with their metadata. despite the name, it returns the full list.

* **Use Case**: "Show me all files in the current workspace."

## System & Maintenance

### `sync_files`

Synchronizes a local directory with the workspace, automatically uploading and embedding supported files.

* **Args**: `source_path` (optional str), `auto_embed` (bool)
* **Use Case**: "Sync the latest code changes to the knowledge base."

### `check_embeddings`

Validates the integrity of embeddings in the workspace, identifying missing or stale vectors.

* **Use Case**: "Verify if all documents are correctly indexed."

### `refresh_embeddings`

Forces a re-index for a specific document, clearing old vectors and re-generating them.

* **Args**: `document_id` (str), `clear_cache` (bool)
* **Use Case**: "The file changed, update its embeddings."