一个真正以「家人」为中心的家庭 AI 操作系统
A Family-First Home AI Operating System
官网：familyclaw.cc | 官方文档：docs.familyclaw.cc | 官方交流群：964063018
FamilyClaw 不是又一个智能音箱 App，也不是一个冷冰冰的设备控制面板。
它是一个懂你全家的 AI 管家 —— 认识每个家庭成员，记住你们的偏好和习惯，在日常生活中默默守护你的家。
传统的智能家居平台以「设备」为中心：灯、空调、传感器。FamilyClaw 把视角翻转过来 —— 家人才是主角，设备只是工具。它认识你、了解你，随着时间推移越来越懂你的家。
让每一个家庭都拥有一位有温度、有记忆、有个性的 AI 管家。 不需要学习复杂的操作，不需要对着音箱喊固定的口令。 只要自然地生活，AI 就能理解你的需求，照顾你的家人。
我们相信，家庭 AI 不应该是科技发烧友的专属玩具，而应该是每一位家庭成员都能舒服使用的贴心助手 —— 无论是忙碌的父母、正在成长的孩子，还是需要关怀的老人。
- 多成员识别：通过声纹（已实现）、人脸（开发中）、蓝牙（计划中）等方式自动识别「现在是谁在说话」
- 成员画像：每个人都有独立的偏好、权限和互动记录
- 角色与权限：为父母、孩子、老人、访客设置不同的使用权限
- 亲属关系：AI 理解家庭成员之间的关系，对话更自然
- 本地大模型深度适配 —— 让 Ollama 等本地模型发挥最大能力，完全离线也能用
- 多角色 AI 助手：管家、营养师、健身教练、学习辅导...... 你的家庭成员各取所需
- 人格化设计：每个 AI 角色有自己的性格、说话风格和服务边界
- 实时对话：基于 WebSocket 的流式响应，和 AI 的对话流畅自然
- 智能提案：AI 会在对话中主动建议行动（比如「要不要帮你设个提醒？」），你确认后自动执行
- 主动关怀：通过计划任务和心跳机制，AI管家会提醒你的会议日程，提醒老人吃药，提醒孩子的网课
这是 FamilyClaw 最核心的能力之一。
- 六种记忆类型：事实、事件、偏好、关系、成长、观察 —— 完整刻画家庭生活
- 越用越懂你：AI 会从日常对话中提取有价值的信息，形成家庭记忆
- 隐私分层：公开、家庭可见、仅自己可见、敏感信息 —— 你的隐私你做主
- 语义搜索：通过向量嵌入实现智能检索，问什么都能找到相关记忆
- Home Assistant 接入：无缝对接 Home Assistant 生态，几千种设备即接即用
- 小米生态：专门适配小米智能家居设备（通过HA进行桥接）
- 对话式控制：不用打开设备面板，直接对 AI 说「把客厅灯调暗一点」
- 设备感知：AI 实时了解所有设备状态，回答更精准
- 智能场景：随着AI管家对你的家庭的了解加深，后续会更加主动的根据家庭的状态、人员的状态智能调节智能家居（开发中）
- 智能回家：识别到家人回家，自动开灯、播报欢迎消息
- 儿童睡前：到了睡觉时间自动调低音量、发送温馨提醒
- 老人关怀：定时问候、用药提醒，让子女安心
- 安全守护：静默时段保护、儿童内容过滤、隐私区域限制
- 对话式创建：聊天中说一句「明天早上 8 点提醒我吃药」，AI 自动创建提醒
- 灵活调度：支持单次、重复、自定义周期
- 升级提醒：重要事项无人响应时自动升级通知方式
- 网页端（H5）：打开浏览器就能用
- iOS / Android：原生移动体验
- 鸿蒙（HarmonyOS）：支持华为生态（开发中）
- 即时通讯接入：钉钉、飞书、企业微信（开发中）、Discord（开发中）、Telegram —— 在你常用的 App 里就能和家庭 AI 对话
FamilyClaw 拥有一套成熟的插件系统，目前内置 32+ 个插件：
| 类型 | 数量 | 示例 |
|---|---|---|
| AI 供应商 | 14 个 | ChatGPT、Claude、Deepseek、通义千问、豆包、Kimi、智谱 GLM、Gemini 等 |
| 通讯渠道 | 6 个 | 钉钉、飞书、企业微信、Discord、Telegram |
| 主题包 | 8 个 | 春和景明、风驰电掣、锦绣前程、星河万里...... |
| 语言包 | 1 个 | 繁体中文 |
| 设备集成 | 2 个 | Home Assistant、小爱音箱 |
| 健康服务 | 1 个 | 基础健康数据 |
支持从插件市场一键安装，也支持 ZIP 手动上传。第三方插件在安全沙箱中运行，不影响主系统稳定性。 支持开发者生态，提供完整的开发文档和规范，开发者可以快速上手进行插件开发，哪怕你是不懂开发的普通人，也可以通过AI来实现快速开发！
本地优先，安全第一 9-14B模型即可稳定运行大部分功能，家用级设备友好，可以做到数据完全本地化。
家人优先，设备其次 不是先买设备再配 App，而是先了解家人再服务生活。
AI 有记忆，不健忘 不同于每次对话都是「初次见面」的通用 AI，FamilyClaw 的 AI 真正记住你的家。
多角色 AI，各司其职 管家处理日常事务，营养师关注饮食健康，健身教练督促锻炼 —— 一个系统，多位专家。
声纹识别，开口即知 不用先解锁手机或报上名字，AI 听到声音就知道是谁。
本地化语音处理 语音识别和声纹比对在本地完成，你的语音数据不会上传到云端。
灵活的 AI 供应商 支持 14 种 AI 供应商，本地 Ollama 和云端 API 随意切换，可设置自动回退策略。
全平台覆盖 一套系统，Web、iOS、Android、鸿蒙全支持。还能在钉钉、飞书、Telegram 里直接使用。
Docker 一键部署 一条命令完成安装，自带数据库、反向代理和进程管理，适合家庭服务器和 NAS。
- 分层长期记忆优化 —— 让 AI 的记忆更智能，重要的事情永远不忘，琐碎的信息自动淡化
- 对话式设备控制升级 —— 更自然的语音/文字设备控制，支持复合指令
- 更多通讯渠道 —— 微信个人号、LINE 等平台的接入
- 成员画像增强 —— 实时感知家庭成员状态，AI 服务更贴心
- 视频流接入 —— 通过接入视频信息，让AI更加了解你的家庭
- 记忆知识图谱 —— 家庭记忆不再是零散的卡片，而是构建成可视化的知识网络
- 插件市场正式上线 —— 完善版本管理和自动更新，社区贡献的插件一键安装
- 智能场景推荐 —— AI 根据家庭生活模式，主动推荐自动化场景
- 多家庭协同 —— 父母和子女的家庭之间建立连接，异地关怀更便捷
- 儿童成长追踪 —— 记录学习进度、兴趣变化，帮助家长更好地了解孩子
- 健康管理进阶 —— 对接健康设备，追踪运动数据和健康指标
这是最简单的安装方式，适合大多数用户。
准备工作：
- 一台能运行 Docker 的设备（家用电脑、NAS、树莓派都行）
- 至少 2GB 内存
安装步骤：
# 1. 如果你已经有docker应用程序，一条命令即可运行
docker run -d \
--name familyclaw \
-p 8080:8080 \
-p 4399:4399 \
-v /srv/familyclaw-data:/data \
jingyi0605/familyclaw:latest
# 2. 等待服务启动（约 60 秒），然后打开浏览器访问
# http://你的设备IP:8080默认安装入口统一使用 latest。只有在需要精确回滚、复现某个旧版本问题，或者明确锁定发布版本时，才手动改成具体标签，例如 jingyi0605/familyclaw:0.1.1。
初始账号：
- 管理员：
user/user（初始化后即失效）
适合想自己折腾的技术用户。
前置要求：
- Python 3.11+
- Node.js 20+
- PostgreSQL 15+
# 1. 克隆项目
git clone https://github.com/your-org/familyclaw.git
cd familyclaw
# 2. 配置数据库
# 创建 PostgreSQL 数据库：familyclaw，用户：familyclaw
# 3. 启动后端
cp apps/api-server/.env.example apps/api-server/.env
# 修改 .env 中的数据库连接信息
cd apps
bash start-api-server.sh
# 脚本会自动创建虚拟环境、安装依赖、运行数据库迁移
# 请在 Git Bash 中执行；Windows 上如果 PATH 里只有 py，脚本会自动回退到 py -3.11
# 4. 启动前端（新开一个终端）
cd .. # 回到项目根目录
npm install --legacy-peer-deps
npm run dev:user-app:h5
# 5. 打开浏览器访问 http://localhost:10086（或终端提示的地址）如果你有小米的小爱音箱，可以启动网关让 AI 管家通过小爱音箱说话：
cd apps
bash start-open-xiaoai-gateway.sh
# 请在 Git Bash 中执行；Windows 上如果 PATH 里只有 py，脚本也会自动识别| 配置项 | 说明 | 默认值 |
|---|---|---|
FAMILYCLAW_DATABASE_URL |
PostgreSQL 数据库地址 | Docker 默认会把最终使用的数据库密码同步到 /data/runtime/secrets/db-password；如需自定义，优先传 FAMILYCLAW_DB_PASSWORD，不要让它和连接串里的密码打架 |
FAMILYCLAW_VOICE_RUNTIME_MODE |
语音模式 | embedded（本地处理） |
FamilyClaw 欢迎社区开发者贡献插件！以下是快速入门。
| 插件类型 | 用途 | 举例 |
|---|---|---|
ai-provider |
接入新的 AI 大模型 | 接入一个新的国产大模型 |
channel |
接入新的通讯平台 | 微信个人号接入 |
integration |
对接外部设备/服务 | 对接扫地机器人 |
action |
添加可执行的操作 | 自动浇花 |
agent-skill |
给 AI 助手添加新技能 | 教 AI 查天气 |
locale-pack |
添加新语言 | 日语支持 |
theme-pack |
添加新主题 | 自定义配色 |
region-provider |
提供地区数据 | 本地天气/地理信息 |
1. 创建插件目录
my-awesome-plugin/
├── manifest.json # 插件描述文件（必须）
├── __init__.py # 入口代码（必须）
├── config_specs.json # 配置表单定义（可选）
└── README.md # 插件说明
2. 编写 manifest.json
{
"id": "my-awesome-plugin",
"name": "我的插件",
"version": "1.0.0",
"api_version": "1",
"types": ["action"],
"description": "一个示例插件",
"entrypoints": {
"action": "my_module:execute"
},
"risk_level": "low"
}3. 编写插件逻辑
async def execute(context, params):
"""插件的主要逻辑"""
# context 包含家庭信息、设备状态等上下文
# params 包含调用参数
return {"status": "ok", "message": "执行成功！"}4. 安装测试
将插件目录压缩为 ZIP，在「插件管理」页面上传安装。第三方插件以子进程模式运行，安装后即刻生效，无需重启服务。
完整的插件开发文档在 docs/开发者文档/插件开发/ 目录下：
| 文档 | 内容 |
|---|---|
| 01 - 插件开发总览 | 架构设计和边界说明 |
| 03 - manifest 字段规范 | manifest.json 完整字段说明 |
| 04 - 目录结构规范 | 推荐的文件组织方式 |
| 06 - 从零开发插件 | 手把手教程 |
| 11 - 配置接入说明 | 动态配置表单开发 |
| 12 - 类型与接口总表 | 所有插件类型的完整接口参考 |
FamilyClaw/
├── apps/
│ ├── api-server/ # Python 后端（FastAPI + SQLAlchemy）
│ │ ├── app/
│ │ │ ├── modules/ # 业务模块（成员、对话、记忆、设备、插件...）
│ │ │ └── plugins/ # 内置插件（32+）
│ │ └── migrations/ # 数据库迁移（Alembic）
│ ├── user-app/ # 跨平台前端（Taro + React + TypeScript）
│ └── open-xiaoai-gateway/ # 设备网关
├── packages/
│ ├── user-core/ # 前端公共库（类型定义、服务层）
│ ├── user-platform/ # 平台适配层
│ └── user-ui/ # UI 组件库
├── docs/ # 项目文档（VitePress 驱动）
├── specs/ # 功能规格书（50+ 篇）
├── docker/ # Docker 部署配置
└── Dockerfile # 一键构建镜像
FamilyClaw 的诞生离不开以下优秀的开源项目：
- FastAPI — 高性能 Python Web 框架，支撑了整个后端 API
- SQLAlchemy — 强大的数据库 ORM
- Taro — 跨平台前端框架，一套代码运行在 Web、iOS、Android、鸿蒙
- React — 用户界面构建库
- Alembic — 数据库迁移管理
- Sherpa-ONNX — 本地语音识别与声纹比对引擎
- Home Assistant — 开源智能家居平台
- open-xiaoai — 让小爱音箱「听见你的声音」，解锁无限可能。
- OpenClaw (MIT License) — 个人 AI 助手项目，FamilyClaw 在通讯渠道插件设计上参考了其优秀的架构理念
- OpenClaw China — 面向中国 IM 平台的 OpenClaw 扩展插件集合，FamilyClaw 在开发钉钉消息渠道时借鉴了此项目
- PostgreSQL — 可靠的关系型数据库
- Nginx — 高性能反向代理
- VitePress — 文档站点生成器
- Docker — 容器化部署
FamilyClaw 是一个充满热情的项目，欢迎各种形式的贡献：
- 提交 Bug 报告或功能建议
- 开发新的插件（AI 供应商、通讯渠道、主题包等）
- 改进文档和翻译
- 分享使用经验
如果你对 FamilyClaw 感兴趣，或者在使用中遇到问题，欢迎通过 Issue 与我们交流。
FamilyClaw is not just another smart home app or a cold device control panel.
It is an AI butler that truly knows your family — it recognizes every family member, remembers your preferences and habits, and quietly takes care of your home in everyday life.
Traditional smart home platforms center around devices: lights, thermostats, sensors. FamilyClaw flips the perspective — your family members are the protagonists, devices are just tools. It knows you, understands you, and grows to know your family better over time.
Give every family a warm, thoughtful, personalized AI butler. No complicated setups. No rigid voice commands. Just live naturally, and the AI understands your needs and cares for your family.
We believe home AI shouldn't be a toy for tech enthusiasts only — it should be a caring assistant that every family member can comfortably use, whether it's busy parents, growing children, or elderly relatives who need extra care.
- Multi-member identification: Automatically recognizes "who's talking" via voiceprint, facial recognition, and Bluetooth
- Individual profiles: Each member has their own preferences, permissions, and interaction history
- Role-based access: Different permission levels for parents, children, elders, and guests
- Family relationships: AI understands family dynamics for more natural conversations
- Deep local model adaptation: Help local models such as Ollama deliver their best performance, with full offline availability
- Multi-persona AI assistants: Butler, Nutritionist, Fitness Coach, Study Coach — each family member gets the help they need
- Personality design: Every AI persona has its own character, speaking style, and service boundaries
- Real-time chat: WebSocket-based streaming responses for smooth, natural conversations
- Smart proposals: AI proactively suggests actions during conversations (e.g., "Shall I set a reminder for that?"), executed after your confirmation
- Proactive care: Through scheduled tasks and heartbeat-style checks, the AI butler can remind you about meetings, remind elders to take medicine, and remind children about online classes
This is one of FamilyClaw's most powerful capabilities.
- Six memory types: Facts, Events, Preferences, Relations, Growth, Observations — a complete portrait of family life
- Gets smarter over time: AI extracts valuable information from daily conversations to build family memory
- Privacy layers: Public, Family-visible, Private, Sensitive — you control your privacy
- Semantic search: Vector embedding-powered smart retrieval finds relevant memories for any query
- Home Assistant integration: Seamless connection with the Home Assistant ecosystem — thousands of devices ready to use
- Xiaomi ecosystem: Purpose-built support for Xiaomi smart home devices through Home Assistant bridging
- Conversational control: Just tell the AI "dim the living room lights" instead of opening a device panel
- Device awareness: AI has real-time knowledge of all device states for more accurate responses
- Smart scenes: As the AI butler learns more about your household and the status of your family members, it can gradually make smarter, more proactive home adjustments (in development)
- Smart Homecoming: Detects family members arriving home, automatically turns on lights and announces a welcome
- Child Bedtime: Automatically lowers volume and sends gentle reminders at bedtime
- Elder Care: Scheduled check-ins and medication reminders to give adult children peace of mind
- Safety Guards: Quiet hours protection, child content filtering, and privacy zone restrictions
- Conversational creation: Say "remind me to take medicine at 8 AM tomorrow" in a chat, and AI creates the reminder automatically
- Flexible scheduling: One-time, recurring, or custom intervals
- Escalation: Automatically increases notification urgency for unacknowledged reminders
- Web (H5): Open a browser and start using it
- iOS / Android: Native mobile experience
- HarmonyOS: Huawei ecosystem support (in development)
- Messaging integrations: DingTalk, Feishu, WeCom (in development), Discord (in development), Telegram — chat with your family AI right in the apps you already use
FamilyClaw features a mature plugin system with 32+ built-in plugins:
| Type | Count | Examples |
|---|---|---|
| AI Providers | 14 | ChatGPT, Claude, Deepseek, Qwen, Doubao, Kimi, GLM, Gemini, etc. |
| Channels | 6 | DingTalk, Feishu, WeCom, Discord, Telegram |
| Themes | 8 | Spring Breeze, Lightning Speed, Starry Night, and more |
| Locale Packs | 1 | Traditional Chinese |
| Device Integrations | 2 | Home Assistant, Xiaomi Speaker |
| Health Services | 1 | Basic health data |
Plugins can be installed with one click from the marketplace or manually uploaded as ZIP files. Third-party plugins run in a secure sandbox without affecting system stability. The project also supports a developer ecosystem with complete documentation and conventions, so contributors can get started quickly. Even non-developers can use AI assistance to build plugins faster.
Local First, Security First Most features can run reliably on 9B-14B models, making FamilyClaw friendly to household-grade hardware while keeping data fully local.
Family First, Devices Second Instead of starting with devices and fitting people around them, FamilyClaw starts with understanding your family and serving their lives.
AI with Memory Unlike generic AI assistants that "meet you for the first time" every conversation, FamilyClaw's AI truly remembers your family.
Multi-Persona AI Team A butler handles daily affairs, a nutritionist watches your diet, a fitness coach keeps you active — one system, multiple experts.
Voiceprint Recognition No need to unlock your phone or introduce yourself — the AI knows who you are just by hearing your voice.
Local Voice Processing Voice recognition and voiceprint matching happen locally. Your voice data never leaves your home.
Flexible AI Providers 14 AI providers supported. Switch freely between local Ollama and cloud APIs, with automatic fallback strategies.
True Cross-Platform One system covering Web, iOS, Android, and HarmonyOS. Plus chat directly through DingTalk, Feishu, or Telegram.
One-Command Docker Deployment A single command installs everything — database, reverse proxy, and process management included. Perfect for home servers and NAS devices.
- Hierarchical memory optimization — Smarter memory management: important things are never forgotten, trivial info fades naturally
- Conversational device control upgrade — More natural voice/text device control with compound commands
- More messaging channels — WeChat personal account, LINE, and more
- Enhanced member profiles — Real-time awareness of family member status for more caring AI service
- Video stream integration — Help the AI understand your household better by bringing in video information
- Memory knowledge graph — Transform scattered memory cards into a visual knowledge network
- Plugin marketplace launch — Version management, auto-updates, and one-click community plugin installation
- Smart scene recommendations — AI proactively suggests automation scenes based on your family's living patterns
- Multi-household collaboration — Connect parents' and children's households for easier long-distance caring
- Child growth tracking — Record learning progress and interest changes to help parents understand their children better
- Advanced health management — Connect health devices, track fitness data and health indicators
The simplest installation method, suitable for most users.
Prerequisites:
- A device that can run Docker (home PC, NAS, Raspberry Pi, etc.)
- At least 2GB of RAM
Steps:
# 1. If you already have Docker installed, one command is enough
docker run -d \
--name familyclaw \
-p 8080:8080 \
-p 4399:4399 \
-v /srv/familyclaw-data:/data \
jingyi0605/familyclaw:latest
# 2. Wait for startup (~60 seconds), then open your browser
# http://your-device-ip:8080Use latest as the default installation entry. Only pin a concrete image tag such as jingyi0605/familyclaw:0.1.1 when you need precise rollback, issue reproduction, or a fixed release target.
Default credentials:
- Admin:
user/user(invalid immediately after initialization)
For tech-savvy users who want to tinker.
Requirements:
- Python 3.11+
- Node.js 20+
- PostgreSQL 15+
# 1. Clone the project
git clone https://github.com/your-org/familyclaw.git
cd familyclaw
# 2. Set up the database
# Create a PostgreSQL database: familyclaw, user: familyclaw
# 3. Start the backend
cp apps/api-server/.env.example apps/api-server/.env
# Edit database connection info in .env
cd apps
bash start-api-server.sh
# The script auto-creates a virtual environment, installs dependencies, and runs migrations
# Run this in Git Bash; on Windows the script also falls back to py -3.11 when python is not on PATH
# 4. Start the frontend (in a new terminal)
cd .. # Back to project root
npm install --legacy-peer-deps
npm run dev:user-app:h5
# 5. Open your browser at http://localhost:10086 (or the address shown in your terminal)If you have a Xiaomi XiaoAI speaker, start the gateway to let the AI butler speak through it:
cd apps
bash start-open-xiaoai-gateway.sh
# Run this in Git Bash; on Windows the script also detects py automatically| Setting | Description | Default |
|---|---|---|
FAMILYCLAW_DATABASE_URL |
PostgreSQL connection string | Docker syncs the effective database password to /data/runtime/secrets/db-password; if you customize it, prefer FAMILYCLAW_DB_PASSWORD and keep it consistent with the password inside the URL |
FAMILYCLAW_VOICE_RUNTIME_MODE |
Voice processing mode | embedded (local) |
FamilyClaw welcomes community developers to contribute plugins! Here's how to get started.
| Plugin Type | Purpose | Example |
|---|---|---|
ai-provider |
Connect new AI models | Add a new LLM provider |
channel |
Connect new messaging platforms | WeChat personal account |
integration |
Interface with external devices/services | Robot vacuum integration |
action |
Add executable operations | Auto-watering plants |
agent-skill |
Add new skills to AI assistants | Teach AI to check weather |
locale-pack |
Add new languages | Japanese support |
theme-pack |
Add new themes | Custom color schemes |
region-provider |
Provide regional data | Local weather/geo info |
1. Create the plugin directory
my-awesome-plugin/
├── manifest.json # Plugin descriptor (required)
├── __init__.py # Entry code (required)
├── config_specs.json # Config form definition (optional)
└── README.md # Plugin documentation
2. Write manifest.json
{
"id": "my-awesome-plugin",
"name": "My Awesome Plugin",
"version": "1.0.0",
"api_version": "1",
"types": ["action"],
"description": "An example plugin",
"entrypoints": {
"action": "my_module:execute"
},
"risk_level": "low"
}3. Write plugin logic
async def execute(context, params):
"""Main plugin logic"""
# context includes family info, device state, etc.
# params includes invocation parameters
return {"status": "ok", "message": "Success!"}4. Install and test
Compress your plugin directory into a ZIP file and upload it via the Plugin Management page. Third-party plugins run in subprocess mode — they take effect immediately without restarting the server.
Complete plugin development docs are available in docs/开发者文档/插件开发/:
| Document | Content |
|---|---|
| 01 - Plugin Development Overview | Architecture and boundary explanation |
| 03 - Manifest Specification | Complete manifest.json field reference |
| 04 - Directory Structure | Recommended file organization |
| 06 - Build a Plugin Walkthrough | Step-by-step tutorial |
| 11 - Configuration Integration | Dynamic config form development |
| 12 - Plugin Types & Contracts | Complete interface reference for all plugin types |
FamilyClaw/
├── apps/
│ ├── api-server/ # Python backend (FastAPI + SQLAlchemy)
│ │ ├── app/
│ │ │ ├── modules/ # Business modules (members, chat, memory, devices, plugins...)
│ │ │ └── plugins/ # Built-in plugins (32+)
│ │ └── migrations/ # Database migrations (Alembic)
│ ├── user-app/ # Cross-platform frontend (Taro + React + TypeScript)
│ └── open-xiaoai-gateway/ # Xiaomi Speaker gateway
├── packages/
│ ├── user-core/ # Shared frontend library (types, services)
│ ├── user-platform/ # Platform adaptation layer
│ └── user-ui/ # UI component library
├── docs/ # Project documentation (powered by VitePress)
├── specs/ # Feature specifications (50+)
├── docker/ # Docker deployment config
└── Dockerfile # One-command image build
FamilyClaw wouldn't exist without these amazing open-source projects:
- FastAPI — High-performance Python web framework powering the entire backend API
- SQLAlchemy — Powerful database ORM
- Taro — Cross-platform frontend framework: one codebase for Web, iOS, Android, and HarmonyOS
- React — User interface library
- Alembic — Database migration management
- Sherpa-ONNX — Local speech recognition and voiceprint matching engine
- Home Assistant — Open-source smart home platform
- open-xiaoai — Lets Xiaomi XiaoAI speakers "hear your voice" and unlock far more possibilities
- OpenClaw (MIT License) — Personal AI assistant project. FamilyClaw drew architectural inspiration from its excellent channel plugin design
- OpenClaw China — OpenClaw extension plugin collection for Chinese IM platforms. FamilyClaw referenced this project when developing the DingTalk messaging channel
- PostgreSQL — Reliable relational database
- Nginx — High-performance reverse proxy
- VitePress — Documentation site generator
- Docker — Containerized deployment
FamilyClaw is a passion-driven project, and we welcome contributions of all kinds:
- Submit bug reports or feature suggestions
- Develop new plugins (AI providers, channels, themes, etc.)
- Improve documentation and translations
- Share your experience
If you're interested in FamilyClaw or need help, feel free to reach out via Issues.
Built with love for every family.
用心，为每一个家。