Getting Started

Upload your first document and search it with natural language.

Prerequisites

RAG Engine is a built-in feature of MCP Gateway Pro. To use it, you need:

  • The AppXen Platform appliance deployed in your AWS account
  • Access to your appliance's console (https://<your-appliance-host>)
1

Open RAG Engine

Sign in to your appliance's console:

  1. Open the stack's ConsoleUrl output. Accept the self-signed certificate warning unless you set a domain.
  2. Sign in as the AdminEmail you gave the stack, using the stack's InstanceId output as the temporary password. The console makes you change it at once.
  3. Check the health panel on the Home page. It shows your AWS identity, the services, and Amazon Bedrock access for each model.

The full steps are in First sign-in. RAG Engine embeds your documents with Amazon Titan Text Embeddings V2, which needs no Anthropic step; the Claude models, used by Agent Orchestrator workflows, do. See Amazon Bedrock model access.

Then navigate to RAG Engine in the sidebar. You will see the Overview page with your knowledge base stats.

If you have not uploaded any documents yet, the stats will show zero sources and zero chunks.

2

Upload a Document

Go to the Sources tab. You can upload documents in two ways:

File Upload

Drag and drop a file onto the upload area, or click to browse. Supports PDF, DOCX, HTML, Markdown, CSV, JSON, plain text, and images (PNG, JPEG, TIFF, BMP).

Text Ingestion

Paste or type text content directly. Give it a filename and click ingest. The text is processed immediately (synchronous).

File uploads are processed asynchronously. The source status will progress through:

queued → processing → ready

If parsing fails, the status changes to failed with an error message.

3

Search Your Knowledge Base

Go to the Search tab. Type a natural language query and hit search. RAG Engine will return the most relevant document chunks ranked by semantic similarity. 

# Example query
What is the refund policy for enterprise customers?

# Returns matching passages from your uploaded documents,
# ranked by relevance with source citations

Adjust the top-k dropdown to control how many results are returned (3, 5, or 10). Your search history is preserved for the session.

4

Use from AI Agents

Any AI agent connected to your MCP Gateway can search your knowledge base using the built-in rag-engine MCP tools. No extra configuration needed. 

curl -X POST https://<your-appliance-host>/mcp \
  -H "Authorization: Bearer axgw_live_k1_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "rag-engine__knowledge_query",
      "arguments": {
        "query": "What is the refund policy?",
        "top_k": 3
      }
    },
    "id": 1
  }'

The gateway returns matching document chunks with citations. Your agent can use these results to answer questions grounded in your documents.

On the aggregated /mcp endpoint tool names are <alias>__<tool>, as above. Connect to /mcp/rag-engine instead and the same tool is just knowledge_query. Until the appliance has a domain it serves a self-signed certificate, which curl rejects unless you add --insecure.

What's Next?