• INVENTORY
  • CONFIG
  • MAPPING
  • HUMAN

How to audit knowledge base resources with your AI

Connect your AI to Amazon Bedrock once. From there, you can pull live configurations for your vector stores and data sources through a simple chat interface.

Ask AI about this page

Short answer

How can I audit my knowledge base with an AI?

Your AI pulls the current settings, model types, and data source links directly from Amazon Bedrock. You get a clear view of your vector stores and mappings without digging through the AWS console. This breakdown shows you exactly what your AI sees.

Where every finding lands.

Audit outcomes

Your AI organizes the technical details it finds into these categories.

INVENTORY

Resource lists

The AI identifies every active knowledge base and data source in your account. You get a clean list of what exists.

CONFIG

Model and vector checks

The AI pulls the specific embedding models and vector store settings being used. This confirms your retrieval setup is correct.

MAPPING

Data source links

The AI checks how your data sources connect to your knowledge bases. You can verify if the right files are being indexed.

HUMAN

Manual fixes

The AI only reads and lists data. If a configuration is wrong, you'll need to go into the AWS console to change it yourself.

The workflow

What your AI does when the request arrives.

The AI uses specific Bedrock tools to fetch your configuration data.

  1. Catalog resources

    The AI scans your account to find all available knowledge bases.

    list_knowledge_bases

  2. Inspect specific bases

    The AI pulls the deep configuration details for a single knowledge base you specify.

    get_knowledge_base

  3. Check data connections

    The AI looks at the individual data sources tied to your knowledge bases.

    list_data_sources

  4. Report findings

    The AI summarizes the models, vector stores, and mappings into a readable format.

    get_knowledge_base

Try it

Copy these to start.

Use these prompts to begin your audit.

Starting points

These are starting points. Swap out the specific ID or name for the one you want to check.

Amazon Bedrock Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_VAqgKgPAMaCVADpiaKW1YIUR9Y3n0l8xHvlUnkn1/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Amazon Bedrock capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "amazon-bedrock-kb-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_VAqgKgPAMaCVADpiaKW1YIUR9Y3n0l8xHvlUnkn1/mcp"
    }
  }
}
Copy into chat04

  • List all my knowledge bases so I can see what's active.

  • Show me the configuration details for knowledge base kb-12345.

  • What data sources are currently mapped to my main knowledge base?

  • Check the embedding model being used in my production knowledge base.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Amazon Bedrock KB once, then ask.

One connection gives your AI access to your Bedrock resources. Your credentials stay encrypted, and you can start auditing immediately.

Connect Amazon Bedrock KB to your AI

FAQ

How this task behaves

  • 01

    Can the AI change my embedding models?

    No. The AI can only read and list your existing configurations. You must make any changes in the AWS console.

  • 02

    Can the AI delete a data source if I find an error?

    No. This MCP is read-only for auditing purposes. It cannot delete or modify any resources.

  • 03

    Does the AI see my actual document content?

    No. The AI only sees the metadata, configurations, and mappings for the knowledge base resources.

  • 04

    How does the AI know which knowledge base to check?

    You tell it the specific ID or name in your prompt, and it uses that to fetch the details.

  • 05

    Can I use this to troubleshoot why a search is failing?

    Yes. By inspecting the embedding models and data source mappings, you can verify if the setup matches your expectations.

  • More questions about Amazon Bedrock KB? The Connector page answers them. See everything the Amazon Bedrock KB Connector can do

Connect Amazon Bedrock to Claude, Cursor, ChatGPT & more