AI Care Assistant for Home-Based Services

Empowering caregivers with AI-driven guidance for safer, smarter, and more personalized home care.

Comment

🧠 AI Care Assistant

Client-side Chrome Extension powered by Gemini Nano and Chrome Built-in AI APIs
基于 Gemini Nano 与 Chrome 内置 AI API 的客户端智能护理助手扩展

💡 Inspiration

EN:
The inspiration for AI Care Assistant came from real-world challenges faced by home-based caregivers — both family members and professionals — who often work without immediate medical guidance. I wanted to build a privacy-safe, offline, intelligent assistant that could provide structured caregiving plans even when the internet or cloud access is unavailable.

ZH:
本项目的灵感来源于居家照护的真实痛点：无论是家庭照护者还是专业护理人员，常常缺乏即时、科学的指导。
我希望打造一个隐私安全、离线可用的智能护理助手，即使在没有网络或云端支持的情况下，也能为照护者提供结构化的护理计划。

💬 What it does

EN:
AI Care Assistant is a Chrome Extension that generates evidence-informed caregiving guidance locally using Chrome’s Built-in AI (Gemini Nano).
When Gemini Nano is unavailable, it automatically switches to a local retrieval-augmented generation (RAG) fallback mode — ensuring continuous, private, and offline access.

ZH:
AI Care Assistant 是一款基于 Chrome 内置 AI（Gemini Nano）的浏览器扩展，能够在本地生成循证护理指导。
当 Gemini Nano 暂不可用时，系统会自动切换为本地检索增强生成（RAG）模式，确保用户在离线和隐私环境下依然能获得完整服务。

🧩 How we built it

EN:

  • Built as a Manifest V3 Chrome Extension using plain HTML, CSS, and JavaScript.
  • Implemented Popup UI for caregiver interaction and Background Service Worker for AI logic.
  • Integrated the Prompt API (Gemini Nano) for on-device text generation.
  • Created a local Markdown knowledge base (care_knowledge_en.md) to support retrieval and fallback generation.
  • Added structured formatting, animations, and visuals for clear, user-friendly caregiving plans.

ZH:

  • 使用 Chrome Manifest V3 构建，前端采用原生 HTML、CSS、JavaScript。
  • 实现了用于交互的 Popup UI 以及负责 AI 调用与逻辑处理的 后台服务线程（Background Service Worker）。
  • 集成 Prompt API（Gemini Nano） 以实现本地文本生成。
  • 构建了一个 Markdown 格式的本地知识库 (care_knowledge_en.md)，支持检索增强与备用生成。
  • 输出结构化内容、配色与动画增强，提升护理计划的可读性与交互体验。

⚙️ Challenges we ran into

EN:
The main challenge was ensuring compatibility across Chrome versions since the Built-in AI APIs are still in Early Preview.
We also faced limitations with Gemini Nano availability on some devices. To overcome this, we designed a progressive fallback system that automatically switches to a local RAG generator when the Prompt API is inaccessible.

ZH:
主要挑战是确保扩展在不同 Chrome 版本间的兼容性，因为内置 AI API 目前仍处于 早期预览阶段。
同时，一些设备暂时无法启用 Gemini Nano。
为此，我们设计了一个 渐进式回退系统（progressive fallback），在 Prompt API 不可用时自动切换到本地 RAG 生成器，实现功能稳定运行。

🏆 Accomplishments that we’re proud of

EN:

  • Successfully built a fully client-side AI system with no server dependency.
  • Achieved privacy-by-design architecture — all processing and generation remain on the device.
  • Developed adaptive output modes for family vs professional caregivers, and brief vs detailed responses.
  • Demonstrated Chrome’s Built-in AI potential for real healthcare applications.

ZH:

  • 成功实现了一个 完全在本地运行的 AI 系统，无需服务器依赖。
  • 建立了 隐私优先的体系架构，所有生成与处理均在设备端完成。
  • 设计了 家庭/专业照护者 及 简版/详版 的自适应输出模式。
  • 展示了 Chrome 内置 AI 在实际健康护理领域中的应用潜力。

📚 What we learned

EN:
We learned how to integrate emerging AI APIs directly into browser environments without relying on cloud services.
We also discovered the importance of fallback strategies for user trust — ensuring a consistent and private user experience regardless of technical limitations.

ZH:
我们学习了如何将新兴的 AI API 直接整合进浏览器环境中，而无需依赖云服务。
同时，我们认识到 回退策略（fallback strategy） 对用户信任的重要性：无论技术限制如何，都要保证一致且安全的体验。

🚀 What's next for AI Care Assistant for Home-Based Services

EN:

  • Integrate additional Chrome Built-in AI APIs such as Summarizer and Proofreader.
  • Enable multilingual support with the Translator API.
  • Expand the local knowledge base for disease-specific guidance.
  • Add daily care plan generation and export features.
  • Explore use cases in home-based long-term care insurance systems.

ZH:

  • 集成更多 Chrome 内置 AI API，如 Summarizer（摘要） 和 Proofreader（校对）。
  • 借助 Translator API 实现 多语言支持。
  • 扩充知识库，涵盖更多疾病与照护场景。
  • 增加 每日护理计划生成 与导出功能。
  • 进一步探索在 居家长期护理保险体系 中的应用潜力。

🧠 Technical Keywords

Chrome Extension · Gemini Nano · Built-in AI APIs · Client-side AI · RAG · Privacy-preserving · Offline AI · Caregiver Support · Geriatric Nursing

Built With

Share this project:

Updates