Set it up
- 1
Turn on AWS Knowledge Base in your AI
In Claude: open Settings → Connectors and add AWS Knowledge Base. In ChatGPT: Settings → Apps & Connectors.
- 2
Add your key
Paste your AWS Knowledge Base API key when your app asks for it — that links your account.
- 3
Just ask
Tell it what you want in plain words — no special commands.
For example, say
“Search my AWS Bedrock Knowledge Base for documentation about our authentication setup”
Get the skill
A ready-made skill that teaches your AI agent to use AWS Knowledge Base well. AWS's official agent toolkit — Bedrock, knowledge bases, core AWS tasks.
npx skills add aws/agent-toolkit-for-aws/skills Quick skills
Copy one, paste it to your AI, watch it work.
Internal Doc Retrieval
“Search our internal onboarding knowledge base for the process to request new AWS IAM roles”
Compliance-Friendly Q&A
“Look up the incident response policy from our compliance KB and summarize the steps”
Multi-KB Routing
“Search the engineering KB for the deploy rollback procedure”
Good to know
- · Bedrock Knowledge Bases cost money per query and per data store — monitor usage
- · IAM permissions need Bedrock read access — misconfigured roles fail silently
- · Indexing is not instant — newly uploaded docs may take minutes to become queryable
When to use
- ✓ You've set up AWS Bedrock Knowledge Bases with your company docs
- ✓ You need RAG (retrieval-augmented generation) over private enterprise content
- ✓ You're building on AWS and want to use Bedrock's managed vector storage
- ✓ You need compliance-friendly AI retrieval within your AWS account
When not to use
- × You don't use AWS or Bedrock (use a different RAG provider)
- × You need public web search — this is for private KB only
- × You haven't indexed your documents into Bedrock yet (that's the prerequisite)
About
Imagine your company has years of policies, runbooks, onboarding guides, and product docs locked away in PDFs and wikis on AWS. Finding the right answer means search-and-scroll until something matches. This server gives your AI assistant a direct line to that material. It connects to Amazon Bedrock Knowledge Bases — AWS's managed search system for your private documents. You ask Claude a plain English question like 'What's our policy for issuing new IAM credentials?' and Claude looks inside your indexed documents, pulls the relevant passages, and writes an answer grounded in what your company actually wrote. The value: nothing leaves your AWS account, retrieval stays under your existing security controls, and you stop hunting through Confluence pages. The catch: someone has to load and index the documents into Bedrock first, and AWS bills per query and per stored vector. Best for teams already running on AWS who want a private 'ask the company' assistant without standing up a separate vector database.
The official AWS Knowledge Base MCP server connects AI assistants to Amazon Bedrock Knowledge Bases. Perform retrieval-augmented generation (RAG) over your enterprise documents, PDFs, and data sources stored in AWS.
Maintained by AWS with enterprise-grade security and scalability. Ideal for organizations that store documentation and knowledge in AWS.
Workflows That Use AWS Knowledge Base MCP
Recommended recipe
Give Claude memory and real documentationPersistent memory across chats, live library docs, and structured reasoning — so answers stop drifting.
Works With
FAQ
What does AWS Knowledge Base MCP do? +
Imagine your company has years of policies, runbooks, onboarding guides, and product docs locked away in PDFs and wikis on AWS. Finding the right answer means search-and-scroll until something matches. This server gives your AI assistant a direct line to that material. It connects to Amazon Bedrock Knowledge Bases — AWS's managed search system for your private documents. You ask Claude a plain English question like 'What's our policy for issuing new IAM credentials?' and Claude looks inside your indexed documents, pulls the relevant passages, and writes an answer grounded in what your company actually wrote. The value: nothing leaves your AWS account, retrieval stays under your existing security controls, and you stop hunting through Confluence pages. The catch: someone has to load and index the documents into Bedrock first, and AWS bills per query and per stored vector. Best for teams already running on AWS who want a private 'ask the company' assistant without standing up a separate vector database.
Do I need to know how to code? +
No. Turn it on in your AI's settings and ask in plain English — no terminal, no coding.
When should I use AWS Knowledge Base MCP? +
Reach for it when you need to: You've set up AWS Bedrock Knowledge Bases with your company docs; You need RAG (retrieval-augmented generation) over private enterprise content; You're building on AWS and want to use Bedrock's managed vector storage; You need compliance-friendly AI retrieval within your AWS account.
When should I avoid AWS Knowledge Base MCP? +
Skip it when: You don't use AWS or Bedrock (use a different RAG provider); You need public web search — this is for private KB only; You haven't indexed your documents into Bedrock yet (that's the prerequisite).
