经 AI Skill Hub 精选评估,机器人AI工作流 获评「强烈推荐」。这款Agent工作流在功能完整性、社区活跃度和易用性方面表现出色,AI 评分 8.0 分,适合有一定技术背景的用户使用。
机器人AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
机器人AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:npm 全局安装 npm install -g robota # 方式二:npx 直接运行(无需安装) npx robota --help # 方式三:项目依赖安装 npm install robota # 方式四:从源码运行 git clone https://github.com/woojubb/robota cd robota npm install npm start
# 命令行使用
robota --help
# 基本用法
robota [options] <input>
# Node.js 代码中使用
const robota = require('robota');
const result = await robota.run(options);
console.log(result);
# robota 配置说明 # 查看配置选项 robota --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量(覆盖配置文件) export ROBOTA_CONFIG="/path/to/config.yml"
Robota is a composable TypeScript library collection for building AI agents — strict types, multi-provider, tool calling, and an extensible plugin/event architecture. You assemble agents from neutral building blocks; @robota-sdk/agent-cli (an AI coding assistant) is a reference app built from these same libraries, not the product itself.
Evaluating with an AI agent? Start at llms.txt — the consumer map (identity, minimal package set, capability matrix, behavior contracts).
Where this is going:VISION.md— Robota builds Robota. The goal is a general development agent capable enough to build even Robota; developing the Robota repo is the validation benchmark (the hardest dogfood), not a Robota-dedicated tool. The capability roadmap lives in.agents/tasks/SELFHOST-*.
npx @robota-sdk/agent-cli
npm install -g @robota-sdk/agent-cli robota
**No Node.js? Install the standalone binary** (macOS / Linux / Windows — nothing else required):
bash
The minimal set is three packages: agent-core + one agent-provider-<vendor> + agent-tools.
import { Robota } from '@robota-sdk/agent-core';
import { AnthropicProvider } from '@robota-sdk/agent-provider-anthropic';
const agent = new Robota({
name: 'MyAgent',
aiProviders: [new AnthropicProvider({ apiKey: process.env.ANTHROPIC_API_KEY })],
defaultModel: {
provider: 'anthropic',
model: 'claude-sonnet-4-6',
},
systemMessage: 'You are a helpful assistant.',
});
const response = await agent.run('Hello!');
Any OpenAI-compatible endpoint works too — AI gateways, Azure, vLLM, local servers — via the OpenAI provider's baseURL (see the quickstart's gateway section).
```bash
Start here (embedding) — the minimal set:
| Package | Description |
|---|---|
[@robota-sdk/agent-core](https://www.npmjs.com/package/@robota-sdk/agent-core) | Robota engine: run/runStream, history, structured output, plugins |
[@robota-sdk/agent-provider-anthropic](https://www.npmjs.com/package/@robota-sdk/agent-provider-anthropic) | Anthropic provider client |
[@robota-sdk/agent-provider-openai](https://www.npmjs.com/package/@robota-sdk/agent-provider-openai) | OpenAI provider client |
[@robota-sdk/agent-provider-openai-compatible](https://www.npmjs.com/package/@robota-sdk/agent-provider-openai-compatible) | OpenAI-compatible clients (DeepSeek, Qwen, Gemma) |
[@robota-sdk/agent-provider-gemini](https://www.npmjs.com/package/@robota-sdk/agent-provider-gemini) | Gemini / Google provider client |
@robota-sdk/agent-provider-bytedance _(not yet published)_ | ByteDance media/video provider client |
[@robota-sdk/agent-tools](https://www.npmjs.com/package/@robota-sdk/agent-tools) | Tool registry, zod-validated function tools, 9 built-in tools |
App assembly — add when you need sessions, permissions, or plugins:
| Package | Description |
|---|---|
[@robota-sdk/agent-framework](https://www.npmjs.com/package/@robota-sdk/agent-framework) | Assembly layer with config/context loading and createQuery() |
[@robota-sdk/agent-session](https://www.npmjs.com/package/@robota-sdk/agent-session) | Session with permissions, hooks, and compaction |
[@robota-sdk/agent-plugin](https://www.npmjs.com/package/@robota-sdk/agent-plugin) | 8 consolidated lifecycle plugins (logging, usage, limits, performance, webhook, …) |
[@robota-sdk/agent-executor](https://www.npmjs.com/package/@robota-sdk/agent-executor) | Execution engine for the agentic loop |
[@robota-sdk/agent-subagent-runner](https://www.npmjs.com/package/@robota-sdk/agent-subagent-runner) | Subagent dispatch runner |
[@robota-sdk/agent-session-analytics](https://www.npmjs.com/package/@robota-sdk/agent-session-analytics) | Session log timing analysis |
@robota-sdk/agent-testing _(internal, not published)_ | Real-PTY E2E test harness |
Products & transports — the reference CLI and its interaction surfaces:
| Package | Description |
|---|---|
[@robota-sdk/agent-cli](https://www.npmjs.com/package/@robota-sdk/agent-cli) | Reference product: interactive terminal AI coding assistant. Self-contained bundle; includes the private DAG/workflow subsystem surfaced as /workflows create "<natural language>" (authors a workflow, runs it, saves it to .workflows/<name>.json) |
[@robota-sdk/agent-command](https://www.npmjs.com/package/@robota-sdk/agent-command) | Slash command modules (/agent, /help, /provider, /preset, /schedule, …) |
[@robota-sdk/agent-preset](https://www.npmjs.com/package/@robota-sdk/agent-preset) | Named agent profiles (preset system) |
[@robota-sdk/agent-interface-transport](https://www.npmjs.com/package/@robota-sdk/agent-interface-transport) | Transport type contracts (zero deps) |
[@robota-sdk/agent-interface-tui](https://www.npmjs.com/package/@robota-sdk/agent-interface-tui) | TUI interaction type contracts (zero deps) |
[@robota-sdk/agent-transport](https://www.npmjs.com/package/@robota-sdk/agent-transport) | Lean transport core (/headless, /testing, /programmatic) |
[@robota-sdk/agent-transport-tui](https://www.npmjs.com/package/@robota-sdk/agent-transport-tui) | TUI transport (Ink/React) |
[@robota-sdk/agent-transport-http](https://www.npmjs.com/package/@robota-sdk/agent-transport-http) | HTTP/REST transport |
[@robota-sdk/agent-transport-ws](https://www.npmjs.com/package/@robota-sdk/agent-transport-ws) | WebSocket transport |
[@robota-sdk/agent-transport-mcp](https://www.npmjs.com/package/@robota-sdk/agent-transport-mcp) | MCP transport |
@robota-sdk/agent-transport-webrtc _(not yet published)_ | P2P remote-control transport |
@robota-sdk/agent-transport-protocol _(not yet published)_ | Wire protocol shared by the transports |
These tables are a curated index, not the full workspace: they name the packages you are most likely to want. They omit most of the 55 workspace-private packages, list one deliberately (agent-testing, marked), and omit several published ones that are not part of a usual assembly. A row marked (not yet published) or (internal, not published) is in the repository but not on the registry. The complete package layout lives in .agents/project-structure.md, which owns it.
高质量的AI工作流框架
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
AI Skill Hub 点评:机器人AI工作流 的核心功能完整,质量优秀。对于自动化工程师和运维人员来说,这是一个值得纳入个人工具库的选择。建议先在非生产环境试用,再逐步推广。
| 原始名称 | robota |
| 原始描述 | 开源AI工作流:TypeScript framework for building AI agents — multi-provider (Anthropic, OpenAI,。⭐17 · TypeScript |
| Topics | AI机器人工作流TypeScript |
| GitHub | https://github.com/woojubb/robota |
| License | MIT |
| 语言 | TypeScript |
收录时间:2026-06-14 · 更新时间:2026-06-16 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端