From the source
Amazon's AWS blog presents SMART, an open source Python SDK and Terraform stack that supplies the WhatsApp channel layer for AI agents already running on Amazon Bedrock AgentCore.
- The post argues the agent is rarely the hard part: inbound messages arrive as deeply nested, type-dependent payloads that must be authenticated and normalized, and media such as voice notes, images and documents each need their own processing pipeline.
- Wiring the channel up otherwise means configuring more than 10 AWS services, and anything sent more than 24 hours after the user's last message must use a pre-approved
Meta message template.
- SMART deploys the whole infrastructure to a customer's AWS account with one command, and the contract with the agent is deliberately small: the agent receives the inbound message and returns text, optionally with rich message attachments.
- AgentCore runtime hosts and scales the agent, while AgentCore memory persists conversation history scoped to each phone number, with a 7-day default retention period.
- Because AgentCore is framework-agnostic, Strands Agents gets memory wired automatically and LangGraph or Pydantic-AI can integrate too.
- Webmotors, originally scoped as a two-month engagement, compressed delivery to six weeks and moved its internal team into controlled production ahead of a full launch.






