Amazon's New AI Remembers Your Past Chats
Your AI assistant could finally remember what you told it last week.
Off-the-shelf AI assistants answer individual questions well but fall short on continuity — a stateless assistant has no idea you mentioned your fast-draining raised beds three weeks ago, that you only use organic fertilizer, or that your petunias were struggling through a heat wave. According to the article, the fix isn't better answers but memory, and the post shows how to build a personal assistant that accumulates context using OpenClaw, an open source agentic system, running on AgentCore runtime, a capability of Amazon Bedrock AgentCore, with AgentCore memory turning disposable chats into durable knowledge.
The running example is Sprout, a gardening assistant, but the architecture is domain-agnostic — swap the persona and skills manifest and the same pipeline serves a support bot, a fitness coach, or an internal help desk. The entire system lives in a single AWS CloudFormation template, deploys with one command, and runs on consumption-based pricing: roughly $1–2 per month for light personal use versus about $35 per month for an always-on Amazon EC2 instance, with those figures given as estimates as of July 2026. Launching your own version requires the Launch Stack button or scripts/deploy.sh, though no build tooling is required to launch.
- Amazon's AgentCore gives AI a long-term memory, so it remembers past conversations.
- It costs about $1–2 per month for light use, much cheaper than an always-on server.
- The system is a do-it-yourself project, not a ready-made app, so it requires some technical setup.
Why It Matters
This could lead to AI assistants that truly know you, saving time and frustration from repeating yourself.