# Amazon Bedrock KB MCP for AI Agents AI Agent Connect

> Amazon Bedrock KB lets your AI agent talk directly to your AWS Bedrock Knowledge Bases. It handles the heavy lifting of semantic search and managed RAG workflows so you don't have to build custom vector pipelines. You can query your internal documents, check ingestion status, and manage your vector stores through a simple chat interface.

## Overview
- **Category:** industry-titans
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_VAqgKgPAMaCVADpiaKW1YIUR9Y3n0l8xHvlUnkn1/ai-agent-connect
- **Tags:** rag, semantic-search, vector-search, foundation-models, data-retrieval, cloud-infrastructure

## Description

Imagine you have a massive library of corporate documents sitting in AWS, but your AI agent can't see them. Usually, you'd have to build a complex pipeline to chunk that data, turn it into embeddings, and host a vector database yourself. This Connector cuts out all that middleman work. It gives your agent a direct line to the Knowledge Bases you already have set up in Amazon Bedrock. You can ask your agent to find specific info in your S3 buckets or summarize internal wikis without moving a single byte of data out of your secure cloud environment. It's a huge win for keeping things compliant while still giving your agent the context it needs to be actually useful. Since Vinkius makes it so easy to swap and manage these connections, you can get your agent up and running with production-grade RAG in minutes. You just point it at your existing AWS setup, and it starts pulling grounded answers from your private data.

## Tools

### list_data_sources
This tool shows you the storage buckets bound to a knowledge base. It helps you audit exactly where your data is coming from.

### list_ingestion_jobs
Use this to see the status of your data syncing operations. It tells you if your documents were successfully chunked and mapped.

### get_knowledge_base
Use this to pull the specific details of a single knowledge base. It's the best way to see individual IDs and settings.

### list_knowledge_bases
This tool lists all your AWS Bedrock knowledge bases. It lets you see your entire collection in one view.

### retrieve
This tool queries your vector index to find relevant text chunks. It provides the exact content and the source URL for each result.

### retrieve_and_generate
Use this to get a full answer grounded in your private documents. It combines retrieval and generation into one simple step.

## Prompt Examples

**Prompt:** 
```
What's the status of my latest document sync?
```

**Response:** 
```
### Ingestion Job Status

| Job ID | Data Source | Status | Documents Processed |
| :--- | :--- | :--- | :--- |
| `XYZ-098` | `Documentation_Bucket` | ✅ **Completed** | 15 |
| `ABC-123` | `HR_Wiki_S3` | ⏳ **In Progress** | 42 |

The job for your Documentation bucket finished successfully at 08:30 today.
```

**Prompt:** 
```
Find the onboarding steps for new hires in our internal wiki.
```

**Response:** 
```
I found the following onboarding steps in your internal wiki:

*   **Day 1:** Complete HR sync for Slack accounts (within 24h).
*   **Day 2:** Hardware issuance and IT protocol review.
*   **Day 3:** Welcome email template distribution.

Source: `s3://hr-docs/onboarding.pdf`
```

**Prompt:** 
```
List all my active knowledge bases and their embedding models.
```

**Response:** 
```
You have the following knowledge bases active in your region:

*   **Internal Wiki KB** (ID: `ABCDE12345`) - Model: `Titan Text v2`
*   **Product Docs KB** (ID: `FGHIJ67890`) - Model: `Cohere Embed English`
```

## Capabilities

### Query internal documents
Search your private vector indexes to find specific text chunks and source URLs.

### Generate grounded answers
Get full responses from your agent that are based strictly on your private datasets.

### Monitor sync progress
Check the real-time status of your document chunking and ingestion pipelines.

### Audit data sources
See which storage buckets and data sources are feeding your knowledge base.

### List vector stores
View all available knowledge bases and embedding models in your specific region.

## Use Cases

### Finding HR policies in large PDFs
An HR manager asks the agent for the remote work policy. The agent uses retrieve to find the exact paragraph in a 200-page PDF and links to the S3 bucket.

### Technical documentation Q&A
A developer asks if a specific API key is in the docs. The agent uses retrieve_and_generate to scan the technical docs and provide a verified answer.

### Sync status monitoring
A cloud admin needs to know if the latest product manual finished syncing. They ask the agent to check the status of the latest ingestion job.

### Audit regional model availability
A researcher needs to see what embedding models are available. They ask the agent to list the knowledge bases and check the assigned models.

## Benefits

- Skip building vector pipelines by using retrieve to query existing AWS indexes directly.
- Get grounded answers instantly with retrieve_and_generate to ensure your agent doesn't hallucinate.
- Monitor your data flow in real time using list_ingestion_jobs to see if your syncs finished.
- Keep your data secure in AWS by using list_data_sources to audit which buckets are connected.
- Manage your regional setup easily by using list_knowledge_bases to see every active store.

## How It Works

The bottom line is you get production-grade RAG capabilities in your chat interface without managing any vector infrastructure.

1. Subscribe to the Amazon Bedrock KB MCP on Vinkius
2. Enter your AWS IAM Role or User Access Credentials
3. Start asking your agent questions about your private AWS data

## Frequently Asked Questions

**Can the Amazon Bedrock KB MCP access my private S3 files?**
Yes, it connects to the Knowledge Bases you've already set up in AWS. If your S3 bucket is already linked as a data source for a knowledge base, your agent can query it.

**Does this Connector help with RAG?**
Yes, it is specifically designed for managed RAG workflows. It allows your agent to retrieve relevant context from your AWS data to generate grounded answers.

**Can I use this with Claude or Cursor?**
Yes, it works with any MCP-compatible client. You can use it to bring your AWS data into Claude, Cursor, Windsurf, or any other compatible agent.

**Do I need to move my data to a new database?**
No, you don't. This Connector works with the Knowledge Bases you already have in AWS. You don't need to migrate your data to a third-party vector database.

**How do I check if my documents are synced?**
You can simply ask your agent to check your ingestion jobs. It will look up the real-time status of your chunking pipelines and tell you if the sync is finished.

**Is my data secure with this Connector?**
Yes, your data stays within your AWS environment. The Connector simply provides a way for your agent to interact with the services you've already authorized in your AWS account.

**Can my AI agent directly run RAG without calling external LLMs?**
Yes! Use the `retrieve_and_generate` capability. Your agent passes the query and a designated Bedrock model ARN. Bedrock handles fetching chunks from the local vector index and synthesizing the final answer inside AWS boundaries, returning a fully grounded response instantly.

**How can I check if new uploaded documents are successfully indexed in my agent?**
Just ask your agent to list ingestion jobs for a specific Knowledge Base ID and Data Source ID. It will report back the exact status (e.g., SYNCING, COMPLETED, FAILED) of chunks being mapped to your vector layout.

**Can I see exactly where an answer came from in my documentation?**
Absolutely. Both the standard `retrieve` functionality and `retrieve_and_generate` calls will parse out the specific origin document URLs (e.g., S3 paths) and expose the exact raw text snippets that mathematically matched your query vector.