AWS publishes full blueprint for a multimodal WhatsApp ordering assistant on Bedrock AgentCore
AWS ML Blog · rss · 2026-09-05
AWS ML Blog published a full architecture tutorial for a WhatsApp ordering assistant built on Amazon Bedrock AgentCore and Amazon Nova 2, spanning text, voice notes, and voice calls over a single WhatsApp Business number.
Key architecture points:
- One shared cross-channel memory (AgentCore memory keyed by hashed customerid) so a customer texting today and calling tomorrow is recognized.
- Amazon Nova 2 Lite handles text; Nova 2 Sonic handles real-time speech. AgentCore Gateway acts as a managed MCP server exposing backend REST APIs as tools (GetMenu, AddToCart, PlaceOrder).
- Each conversation runs in an isolated microVM; only the voice-call runtime needs a VPC.
- Inbound webhooks flow through API Gateway + Lambda + SQS with instant 200 acks; channel layer is decoupled from ordering logic.
- Everything deployed via AWS CDK with DynamoDB, Location Service, KVS TURN relay, Secrets Manager, and more.
A solid reference for production-grade agent architecture (MCP, cross-channel memory, multimodal voice).
More from coding & agent
- GLM-5.3 launches with vision on Baseten, Terminal-Bench 3.0 jumps 4.6% to 28.3% — baseten · 2026-09-05
- Swarms v15 rebuilds multi-agent execution with 10+ MCP tutorials — KyeGomezB · 2026-09-05
- Google GenAI SDK for Kotlin hits 1.0: idiomatic multiplatform access to Gemini — rseroter · 2026-09-05
- Cross-Model Code Review: Having Claude and Copilot CLI Battle Over Refactoring — DanWahlin · 2026-09-05
- Developer Uses Claude Code to Ship a Working F-Zero X Port to 3DS at Near 60fps — killermike523 · 2026-09-05
- Coinbase's x402 protocol replaces 700+ API keys with a single wallet signature for AI agents — kleffew94 · 2026-09-05