# GroundX MCP for AI Agents AI Agent Connect

> GroundX is a data search and RAG optimization platform. It lets your AI agent crawl websites, ingest enterprise documents into structured buckets, and perform semantic searches across your private data. It handles the heavy lifting of chunking and indexing so your agent actually finds what you need.

## Overview
- **Category:** knowledge-management
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_mAVynUnnXcCNvXWpgEW86XlUhDZSOLIyZ2eNgeMy/ai-agent-connect
- **Tags:** rag-as-a-service, data-search, document-retrieval, enterprise-data, semantic-indexing, llm-context

## Description

GroundX lets you give your AI agent a reliable memory. Instead of throwing a bunch of files into a prompt and hoping it remembers the details, you can organize your data into buckets and groups. This Connector handles the heavy lifting of indexing your PDFs, websites, and local files so your agent can find exactly what it needs when you ask a question. It makes RAG pipelines actually useful because it focuses on retrieving relevant chunks rather than just dumping everything into the context window. You can see your data structure, check if your imports are finished, and manage your workflows in one place. It is a much cleaner way to handle enterprise data than trying to manage everything manually. You can find this in the Vinkius catalog, which hosts over 4,000 Connectors, letting you connect once from your preferred client to access the entire library.

## Tools

### create_bucket
Set up a new container for a specific set of data or project. This keeps your different data types from getting mixed up.

### create_group
Make a new category to group your buckets more logically. It helps organize large amounts of data into manageable themes.

### get_customer_info
Pull account details and customer information from your records. This lets your agent provide personalized answers based on user data.

### get_ingest_status
Check if your latest data import is finished or still processing. This ensures your agent does not try to search for data that is not ready yet.

### ingest_documents
Upload files from local paths or URLs into your GroundX repository. It is the best way to move static files into your searchable index.

### ingest_website
Automatically crawl and index every page on a specific website. This keeps your agent updated on external web resources without manual updates.

### list_buckets
See all the different containers where your documents are stored. Use this to get a high-level view of your data organization.

### list_content
View a list of every document you have already ingested. This helps you confirm that your files are actually in the system.

### list_groups
Browse the various categories you have created to group your buckets. It helps organize large amounts of data into manageable themes.

### list_workflows
See all the active RAG workflows currently running in your account. Use this to monitor your active data pipelines.

### search_content
Run a semantic search to find relevant information across your entire data set. This is how your agent finds the best answers to your questions.

### search_documents
Find specific files based on their metadata or the text inside them. Use this when you need to narrow down your search to specific files.

## Prompt Examples

**Prompt:** 
```
What is our current refund policy for international orders?
```

**Response:** 
```
### International Refund Policy

Based on the **Support Docs**, here is the current policy:

* **Processing Time**: 5-7 business days
* **Requirement**: Must be requested via the customer portal
* **Conditions**: 
    * Full refund for damaged goods
    * Partial refund for changed minds

*Note: This applies only to orders placed in the last 30 days.*
```

**Prompt:** 
```
Show me the list of buckets I have set up.
```

**Response:** 
```
You currently have the following buckets active:

| Bucket Name | ID | Description |
| :--- | :--- | :--- |
| **Knowledge Base** | 101 | Internal company wiki and HR docs |
| **Support Docs** | 102 | Customer facing support manuals |
```

**Prompt:** 
```
Is the website ingestion still running?
```

**Response:** 
```
The ingestion task for the **Product Specs Website** is currently: 🟢 **Processing**

* **Progress**: 64%
* **Status**: Crawling remaining pages...
* **Estimated Completion**: 3 minutes
```

## Capabilities

### Crawl and ingest entire websites
Automatically turn a website URL into a searchable knowledge base for your agent.

### Organize documents into buckets
Group your data into logical containers to keep different topics separated.

### Perform semantic searches
Let your agent find the most relevant information across all your indexed content.

### Track ingestion status
Check if your latest data imports are finished and ready for your agent to use.

### Retrieve specific documents
Pull out exact files based on metadata or content when you need specific details.

## Use Cases

### Searching internal support manuals
A support lead asks the agent about a refund policy. The agent uses search_documents to find the exact policy in the 'Support' bucket.

### Updating agent knowledge from a website
A product manager wants the agent to know the latest specs. They use ingest_website to crawl the company's public product page.

### Querying a company wiki
An employee asks about vacation days. The agent uses search_content to pull the relevant paragraphs from the HR bucket.

### Monitoring data pipelines
An engineer uses list_workflows to check the status of various data ingestion tasks across the company.

## Benefits

- Stop manually copying and pasting text into your prompts by using ingest_documents to build a permanent knowledge base.
- Get accurate answers from your agent because search_content uses semantic indexing to find the most relevant data chunks.
- Keep your data organized by using create_bucket and create_group to separate internal wikis from customer support docs.
- Save time on data collection by using ingest_website to automatically crawl and index entire web resources.
- Stay on top of your data pipeline by checking get_ingest_status to see when your latest files are ready for use.

## How It Works

The bottom line is it turns your static files into a dynamic, searchable brain for your AI agent.

1. Create buckets for your different data types like support docs or internal wikis.
2. Ingest your files or websites into those buckets using the ingestion tools.
3. Ask your agent to search the content and it will pull the best matches automatically.

## Frequently Asked Questions

**How does GroundX help my AI find info?**
GroundX indexes your private files so your AI agent can perform a semantic search. Instead of reading every file, the agent only pulls the most relevant chunks of information to answer your specific question.

**Can I use GroundX for my internal wiki?**
Yes, that is one of the primary use cases. You can ingest your wiki pages and organize them into buckets so your agent can act as an internal knowledge expert.

**Does it work with my existing PDFs?**
Yes, you can upload local PDFs using the ingestion tools. GroundX will handle the indexing so your agent can query the content of those documents easily.

**How do I keep my data organized?**
You can create different buckets for different topics, like 'HR', 'Technical', or 'Sales'. You can also create groups to organize those buckets into broader categories.

**Is it hard to set up GroundX?**
It is designed to be straightforward. You connect your account, create buckets for your data, and start ingesting your sources. Your AI agent handles the rest of the retrieval logic.

**Can it search my company website?**
Yes, the tool includes a specific function to crawl and ingest content from a website URL, turning your public-facing site into a searchable knowledge base for your agent.

**How do I query my indexed documents?**
Simply ask the AI agent to search for a specific term or concept, and it will query the GroundX API to retrieve the most relevant textual chunks.

**Can I manage data buckets from the agent?**
Yes, you can list your active buckets, check their document count, and verify index status.

**Does it support adding new files to a bucket?**
Currently, the integration focuses on querying the optimized indexes. File ingestion should be managed through the GroundX dashboard or a separate pipeline.