This repository contains sample applications built using the Fullstack AgentCore Solution Template (FAST) as a starting point. Each sample demonstrates how to customize FAST for different use cases while leveraging AWS AgentCore.
⚠️ Important: These samples are not production-ready. They pass automated security scanning at the time of contribution but are not guaranteed to receive ongoing security patches or dependency updates. You must thoroughly review any sample code before deploying to production. See SECURITY.md for details.
While FAST provides a fully functional out-of-the-box chat application, it's designed to be customized for any use case that leverages AgentCore. These samples serve as:
- Starting points for similar projects
- A diverse set of examples of how others have extended FAST for a variety of use cases
- Learning resources for engineers
| Sample | Description |
|---|---|
| Restaurant Assistant | Knowledge base integration, reservation management, and customer-facing chat widget |
| CopilotKit Generative UI | Generative UI, shared state, and human-in-the-loop interactions via CopilotKit |
| LLM Council | An implementation of "Council of LLMs" pattern on AWS. Builds consensus among multiple diverse LLMs. |
| Dual Monitoring System | Dual-layer monitoring for agentic solutions using AgentCore Evaluations and AWS DevOps Agent |
| AgentCore AWS Specialist Agent | AWS specialist chat agent with Gateway MCP tools, long-term memory, web search, Skills on Runtime, and a NAT-free VPC |
Description: A restaurant assistant application with knowledge base integration, reservation management, and a professional customer-facing interface.
Built on FAST: v0.4.1
Key Differences from FAST: Adds an s3-vector backed knowledge base, DynamoDB reservations table, custom reservation tools, restaurant-themed landing page with chat widget, and file upload capabilities
Use Case: Building customer service assistants for hospitality businesses or any domain requiring knowledge base integration with transactional capabilities
Description: Adds CopilotKit as the frontend framework on top of FAST, enabling generative UI (inline charts and components rendered from tool calls), bidirectional shared state between the agent and UI, and human-in-the-loop interactions.
Built on FAST: v0.4.1
Key Differences from FAST: Replaces the baseline frontend with CopilotKit, adds a CopilotKit Runtime Lambda as a server-side bridge to AgentCore, and includes both LangGraph and Strands agent patterns with CopilotKit middleware.
Use Case: Building agent-native applications where the AI drives the UI — not just chat — including dashboards, collaborative canvases, and interactive workflows.
Description: An implementation of "Council of LLMs" pattern on AWS. Multiple diverse LLMs collaborate through a 3-stage deliberation process -- independent responses, anonymized peer ranking, and chairman synthesis -- to produce higher-quality answers than any single model alone.
Key Differences from FAST: Replaces single-agent pattern with multi-model council orchestration, parallel Bedrock Converse API invocations across 4 providers (Anthropic, Meta, Amazon, Cohere), anonymized peer ranking with aggregate scoring, chairman synthesis stage, custom streaming event format with stage-by-stage SSE updates, council-specific React UI with tabbed model responses and ranking matrix
Use Case: Building applications where response quality matters more than latency, reducing single-model bias, combining strengths of diverse model providers, or any use case benefiting from collaborative AI deliberation
Description: A dual-layer monitoring architecture for agentic solutions in dev/prod env. Layer 1 uses AgentCore Evaluations to continuously score live agent interactions with LLM-as-a-Judge methodology, surfacing quality degradation that infrastructure metrics miss. Layer 2 uses AWS DevOps Agent to autonomously investigate infrastructure incidents — pulling CloudWatch logs, building resource topology graphs, and tracing complete failure paths with remediation steps.
Built on FAST: v0.4.1
Key Differences from FAST: Adds an evaluation dashboard (session explorer, on-demand evaluation, AI pattern analysis, prompt improvement, online evaluation config), a DevOps Agent incident submission interface with SigV4-signed webhook proxy, a four-agent Strands swarm as the demo workload (supervisor, flight, user, and reservation agents), and additional CDK stacks for evaluation and DevOps Agent infrastructure.
Use Case: Monitoring multi-agent systems in production where traditional infrastructure metrics are insufficient — particularly swarm-based architectures with dynamic handoffs where quality degradation and failure propagation are hard to detect.
Description: An AWS specialist chat agent showcased at the AgentCore booth at AWS Summit Japan 2026. The agent reasons about AWS, calls AWS APIs and managed tools through AgentCore Gateway, searches the web, executes code, and remembers facts across sessions.
Built on FAST: v0.4.1
Key Differences from FAST: Adds selectable models (Claude and OpenAI GPT on Bedrock via a CDK model registry), AgentCore Memory long-term memory with an LTM-listing MCP server, the Amazon-managed Web Search connector, a chat-history sidebar (API Gateway + Lambda + DynamoDB), a fully closed NAT-free VPC using only VPC endpoints, AWS Skills mounted from S3 Files at /mnt/skills, speculative pre-warming to cut cold-start latency, and multi-MCP-server management gated per user department by Cedar ABAC.
Use Case: Building production-oriented specialist agents that need fine-grained per-user tool authorization, private networking, long-term memory, and multiple MCP tool sources on AgentCore.
Have you built something with FAST? We'd love to see it! Please see CONTRIBUTING.md for guidelines on how to contribute your sample application.
For questions about:
- FAST itself: See the main FAST repository
- Specific samples: Open an issue in this repository
- Contributing samples: See CONTRIBUTING.md
This project is licensed under the Apache-2.0 License.