restaurant-assistant

Restaurant Assistant Sample

Built on the Fullstack AgentCore Solution Template (FAST) v0.4.1.

This sample demonstrates how to transform the FAST baseline chat application into a restaurant assistant with knowledge base integration, reservation management, and a professional customer-facing interface.

Major Additions to FAST:

• Separate Knowledge Base stack with S3 Vectors
• DynamoDB reservations table with composite key (booking_id, restaurant_name)
• Custom reservation tools Lambda with tool name routing pattern
• Restaurant-themed landing page with floating chat widget
• File upload with presigned URLs and Knowledge Base sync endpoint
• Restaurant Helper agent persona with Knowledge Base retrieval

Major Removals from FAST:

• Sample tool Lambda (replaced with reservation tools)
• Code Interpreter tools (replaced with KB retrieve)
• Default chat-only interface (replaced with landing page + widget)
• Unused agent patterns (langgraph, claude-agent-sdk)
• Terraform infrastructure (CDK only)

Deployment

cd infra-cdk
npm install
cdk bootstrap  # once per account/region
cdk deploy fast-restaurant-kb
cdk deploy fast-restaurant
cd ..
python scripts/deploy-frontend.py

After deployment, upload the .docx files from sample-data/ via the "Manage Docs" button in the UI, then click "Sync Knowledge Base."

Application Screenshot

Architecture

The architecture follows the FAST baseline with these additions:

• Knowledge Base Stack (fast-restaurant-kb): S3 data bucket, S3 Vector Bucket + Index, Bedrock Knowledge Base with S3 Vectors storage, S3 Data Source
• Backend additions: DynamoDB reservations table, reservation tools Lambda behind AgentCore Gateway, presigned URL Lambda, sync KB Lambda, /upload and /sync API Gateway endpoints
• Agent: Strands agent with Restaurant Helper persona, retrieve tool for KB queries, gateway MCP tools for reservations
• Frontend: Restaurant landing page with hero section, featured restaurants, floating chat widget, and KB document upload panel

Sample Queries

After uploading the sample data and syncing the Knowledge Base, try these queries in the chat widget:

• "What restaurants are available?"
• "Tell me about the best French restaurant"
• "Make a reservation for 2 at Bistro Parisienne on Friday at 7pm under the name John"
• "Check on the status of my reservation with booking ID booking-abc123 at Bistro Parisienne"
• "Cancel my reservation booking-abc123 at Bistro Parisienne"
• "What's on the menu at Ember?"
• "Which restaurants are good for a large group?"

Sample Data

The .docx files in sample-data/ contain fictional restaurant descriptions, menus, and locations. The restaurant images on the landing page are AI-generated and do not depict real establishments.

Security

This asset represents a proof-of-value and is not intended as a production-ready solution. See the FAST documentation for security best practices.

License

This project is licensed under the Apache-2.0 License.