nanobot是一款超轻量、自托管的个人AI代理，适用于工作区、频道、自动化及长期工作流。
简介
nanobot 是一款超轻量、自托管的个人 AI 代理，专为追求高效与隐私的用户打造。它摒弃了传统 AI 工具的臃肿架构，以极低的资源占用提供强大的智能服务。无论是嵌入工作区、接入频道、执行自动化任务，还是支撑长期工作流，nanobot 都能以灵活、安全的方式成为您的数字助手。自托管特性确保您的数据完全由您掌控，适用于对隐私和定制化有高要求的个人及团队。
主要功能
- 智能对话与任务执行：支持自然语言交互，快速响应查询并完成指定操作，如信息检索、日程安排等。
- 工作区集成：可无缝嵌入 Slack、Discord 等协作平台，在频道中直接调用 AI 能力，提升团队效率。
- 自动化流程：通过可配置的触发器与动作，实现邮件自动回复、数据抓取、通知推送等重复性任务的自动化。
- 长期工作流管理：支持多步骤、依赖型任务的编排与持久化执行，适合复杂的项目跟进与数据处理场景。
- 轻量化部署：基于高效架构设计，可在低配服务器或树莓派上流畅运行，启动速度快，内存占用低。
特色优势
- 极致轻量：核心程序体积小于 10MB，运行时内存占用仅 50MB 左右，是同类产品中资源消耗最低的解决方案之一。
- 完全自托管：所有数据与模型运行于您自己的服务器，杜绝第三方数据泄露风险，满足企业级隐私合规要求。
- 高度可定制：开放 API 与插件系统，允许开发者根据需求扩展功能、接入自定义模型或修改行为逻辑。
- 离线可用：支持完全离线模式，无需联网即可完成基础任务，适合内网环境或对网络依赖敏感的场景。
- 零配置启动：提供一键部署脚本与 Docker 镜像，从下载到运行仅需 2 分钟，降低使用门槛。
适用人群
- 个人开发者与技术爱好者：希望快速搭建私有 AI 助手，用于学习、实验或日常自动化。
- 远程团队与小型企业：需要在不暴露敏感数据的前提下，为协作工具（如飞书、企业微信）注入 AI 能力。
- 运维与自动化工程师：寻求一个稳定的工作流引擎，用于监控、告警、日志分析等后台任务。
- 隐私敏感用户：对数据主权有严格要求，不愿将个人或公司信息交由云端服务处理。
常见问题
Q: nanobot 需要什么样的硬件配置？
A: 推荐最低 1 核 CPU、512MB 内存、5GB 磁盘空间。实际运行中，空闲时内存占用约 50MB，负载时不超过 200MB。
Q: 是否支持接入 GPT、Claude 等外部模型？
A: 支持。nanobot 设计了统一的模型接口，您可以通过配置文件切换至 OpenAI、Anthropic 或本地开源的 LLM 模型。
Q: 如何保证数据安全？
A: 所有数据存储于您自己的服务器，nanobot 不会向外部发送任何未经用户授权的信息。您可以启用 HTTPS 和访问密钥来进一步增强安全性。
Q: 有无社区或技术支持？
A: 项目在 GitHub 上开源，提供详细的文档与示例。您也可以通过官方论坛或 Discord 频道与其他用户交流、获取帮助。
English
Introduction
nanobot is an ultra-lightweight, self-hosted personal AI agent designed for users who value efficiency and privacy. It does away with the bloated architecture of traditional AI tools, delivering powerful intelligent services with minimal resource usage. Whether embedded in a workspace, connected to channels, executing automation tasks, or supporting long-running workflows, nanobot serves as your flexible and secure digital assistant. Its self-hosted nature ensures your data remains fully under your control, making it ideal for individuals and teams with high demands for privacy and customization.
Key Features
- Intelligent conversations and task execution: Supports natural language interaction, quickly responding to queries and carrying out specified actions such as information retrieval, scheduling, and more.
- Workspace integration: Seamlessly integrates with collaboration platforms like Slack and Discord, bringing AI capabilities directly into channels to boost team productivity.
- Automation workflows: Uses configurable triggers and actions to automate repetitive tasks such as email auto-replies, data scraping, and notification pushes.
- Long-running workflow management: Supports orchestration and persistent execution of multi-step, dependent tasks, suitable for complex project tracking and data processing scenarios.
- Lightweight deployment: Built on an efficient architecture, it runs smoothly on low-spec servers or Raspberry Pi, with fast startup and low memory usage.
Highlights
- Extremely lightweight: The core program is under 10MB, with a runtime memory footprint of only around 50MB, making it one of the lowest-resource solutions in its category.
- Fully self-hosted: All data and models run on your own server, eliminating the risk of third-party data leakage and meeting enterprise-grade privacy compliance requirements.
- Highly customizable: An open API and plugin system lets developers extend functionality, integrate custom models, or modify behavior logic as needed.
- Offline capable: Supports a fully offline mode, allowing basic tasks to be completed without an internet connection, ideal for intranet environments or scenarios sensitive to network dependency.
- Zero-configuration startup: Offers one-click deployment scripts and Docker images, taking only 2 minutes from download to running, lowering the barrier to entry.
Who It's For
- Individual developers and tech enthusiasts: Those looking to quickly set up a private AI assistant for learning, experimentation, or everyday automation.
- Remote teams and small businesses: Those who need to add AI capabilities to collaboration tools (such as Feishu or WeChat Work) without exposing sensitive data.
- Ops and automation engineers: Those seeking a stable workflow
