Orca 工作流
经 AI Skill Hub 精选评估，Orca 工作流 获评「推荐使用」。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 7.5 分，适合有一定技术背景的用户使用。
📚 深度解析
Orca 工作流 工作流的设计遵循"最小配置，最大复用"原则：核心逻辑已经封装好，用户只需配置自己的 API Key 和业务参数即可快速上手。工作流内置错误处理和重试机制，在网络波动或 API 限速等情况下仍能稳定运行，适合作为生产环境的自动化基础设施。
在实际部署时，建议先在测试环境中运行 3-5 次，验证各个环节的输出结果符合预期，再部署到生产环境。AI Skill Hub 评分 7.5 分，是同类 Agent 工作流中的精选推荐。
📋 工具概览
Orca 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。
📖 中文文档
Orca 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。
- 可视化 Agent 工作流编排，无需编写复杂代码
- 支持多步骤自动化任务链，实现全流程无人值守
- 与外部 API、数据库和第三方服务无缝集成
- 内置错误处理与自动重试机制，保障稳定运行
- 提供可复用的自动化模板，快速在同类场景部署
- 自动化日常重复性工作，将精力集中于创造性任务
- 构建数据采集 → 处理 → 输出的完整自动化管线
- 实现跨平台、跨系统的数据流转和业务协同
# 方式一：npm 全局安装 npm install -g orca-lang # 方式二：npx 直接运行（无需安装） npx orca-lang --help # 方式三：项目依赖安装 npm install orca-lang # 方式四：从源码运行 git clone https://github.com/jascal/orca-lang cd orca-lang npm install npm start
- 访问 GitHub 仓库获取工作流文件
- 在对应平台（Dify / Flowise / Make 等）中找到「导入工作流」功能
- 上传工作流文件
- 按照提示配置必要的环境变量和 API Key
- 运行测试确认流程正常后投入使用
# 命令行使用
orca-lang --help
# 基本用法
orca-lang [options] <input>
# Node.js 代码中使用
const orca_lang = require('orca-lang');
const result = await orca_lang.run(options);
console.log(result);
# orca-lang 配置说明 # 查看配置选项 orca-lang --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 ORCA_LANG_CONFIG="/path/to/config.yml"
Orca
Orchestrated State Machine Language — a two-layer architecture for reliable LLM code generation.
The core insight: LLMs generate flat transition tables reliably, but they struggle to guarantee topology correctness on their own. Orca separates program structure (state machine topology) from computation (action functions), then verifies the structure automatically before any code runs.
Machines are written in plain Markdown — a format LLMs can read and write natively.
---
Features
Language - States with [initial] / [final] markers, descriptions, on_entry / on_exit actions - Transitions as a flat table — the format LLMs generate most reliably - Guard expressions: comparisons, boolean logic, null checks - Hierarchical (nested) states - Parallel regions with all-final / any-final / custom sync strategies - Timeouts: timeout: 30s -> state_name - Ignored events: ignore: EVENT_NAME - Machine invocation: one machine calling another, with input mapping and completion events - Multi-machine files: multiple machines in one .orca.md separated by --- - ## effects section: named I/O schemas for external side effects - Decision tables: co-located conditional logic without guard explosion — verified for completeness, consistency, and cross-machine reachability
Verifier - Reachability: every state is reachable from the initial state - Deadlock detection: every non-final state has an outgoing transition - Completeness: every (state, event) pair is handled or explicitly ignored - Guard determinism: multi-transition guards are mutually exclusive - Property checking: bounded model checking with BFS — reachable, unreachable, passes_through, live, responds, invariant - Cross-machine: cycle detection, child reachability to final state, input mapping validation - Effect consistency: ORPHAN_EFFECT (declared but unused) and UNDECLARED_EFFECT (referenced but not declared) - Decision table checks: completeness, consistency, redundancy, coverage gap, dead guards, DT-constrained reachability — see DECISION_TABLES.md
Compilers - XState v5 createMachine() config - Mermaid stateDiagram-v2
Runtimes (standalone — no XState dependency) - TypeScript (@orcalang/orca-runtime-ts) - Python (orca-runtime-python) - Go (orca-runtime-go)
All three runtimes share the same feature set: guard evaluation, action registration, event bus (pub/sub + request/response), timeouts, parallel regions, snapshot/restore, machine invocation, persistence, and structured logging.
---
Language features
Python packages (runtime + demos, requires Python >= 3.11)
pnpm run setup:python
nanoGPT training orchestrator (Python, no torch required for tests)
pnpm run test:demo-nanolab
Setup
```bash
Compile to XState
npx tsx src/index.ts compile xstate examples/payment-processor.orca.md
Compile to Mermaid
npx tsx src/index.ts compile mermaid examples/text-adventure.orca.md
Install torch with GPU support, then run the full pipeline
.venv/bin/pip install torch torchvision torchaudio numpy requests pnpm run run:demo-nanolab ```
---
Skills & MCP setup
Orca ships six Claude Code skills backed by the @orcalang/orca-mcp-server MCP server. The skills call MCP tools directly — no shell or file access needed.
| Skill | Trigger | What it does |
|---|---|---|
/orca-generate | <spec> | Generate a verified machine from a natural language spec |
/orca-generate-multi | <spec> | Generate a coordinated multi-machine system |
/orca-verify | [file] | Verify a machine for errors and warnings |
/orca-refine | [file] | Auto-fix verification errors using an LLM |
/orca-compile | [xstate\|mermaid] [file] | Compile to XState TypeScript or Mermaid |
/orca-actions | [typescript\|python\|go] [file] | Generate action scaffold stubs |
Skills that use an LLM (/orca-generate, /orca-generate-multi, /orca-refine, and optionally /orca-actions --use-llm) call the MCP server, which calls your configured LLM provider. Skills that are purely structural (/orca-verify, /orca-compile, plain /orca-actions) never make LLM calls.
Examples
All in packages/orca-lang/examples/:
| File | Description |
|---|---|
simple-toggle.orca.md | Minimal 2-state machine |
payment-processor.orca.md | Guards, retries, effects |
text-adventure.orca.md | Multi-state game engine |
hierarchical-game.orca.md | Nested compound states |
parallel-order.orca.md | Parallel regions with sync |
payment-with-properties.orca.md | Bounded model checking properties |
key-exchange.orca.md | Multi-machine: client/server key exchange protocol |
invocation-order.orca.md | Multi-machine: order processing with child invocations |
saas-auth.orca.md | SaaS authentication and registration flow |
health-check.orca.md | Health check machine used by the dogfood runner |
simple-discount.orca.md | Minimal standalone decision table |
payment-routing.orca.md | Payment gateway router decision table |
shipping-rules.orca.md | Shipping cost calculator decision table |
payment-with-routing.orca.md | Combined machine + decision table |
See DECISION_TABLES.md for a full guide to decision tables.
---
Running the demos
```bash
MCP server environment variables
| Variable | Required | Description |
|---|---|---|
ORCA_API_KEY | Yes | API key for your LLM provider |
ORCA_PROVIDER | Yes | anthropic, openai, ollama, or grok |
ORCA_BASE_URL | No | Override the provider's default base URL (for OpenAI-compatible APIs) |
ORCA_MODEL | No | Model name (defaults to claude-sonnet-4-6 for Anthropic) |
Use ORCA_PROVIDER=openai with ORCA_BASE_URL for any OpenAI-compatible provider (MiniMax, Together, local vLLM, etc.).
---
TypeScript packages
pnpm install pnpm build
Go packages
pnpm run setup:go pnpm run build:demo-go ```
---
All TypeScript packages
pnpm test
or from the package directory:
cd packages/mcp-server && npx tsc
Create `.mcp.json` at the project root (it is already in `.gitignore`), then restart Claude Code. Skills are auto-discovered from `.claude/skills/` — no additional configuration needed.
> **Node.js version** — Claude Code may use an older Node.js than the Node 18+ required by this package (ESM). If the server fails to start, add a `PATH` entry to `env` that puts your system Node's `bin` directory first:
> dirname $(which npx) to find the correct path. On nvm it will be something like ~/.nvm/versions/node/v22.x.x/bin.
---
Orca 项目简介：Orca 是一个用于构建状态机的语言，支持多种特性，如状态标记、转换表、守卫表达式等。Orca 可以与 LLMs（大语言模型）集成，用于训练和部署状态机。
Orca 的功能特点包括：支持状态标记、转换表、守卫表达式、层次化状态、并行区域等。Orca 还支持与 LLMs 的集成，用于训练和部署状态机。
Orca 需要 Python >= 3.11 的环境，需要安装相关的 Python 包和依赖项。
安装 Orca 可以使用以下步骤：使用 `pnpm run setup:python` 安装 Python 环境，使用 `npx tsx src/index.ts compile xstate` 或 `npx tsx src/index.ts compile mermaid` 编译 Orca 代码。
Orca 提供了多个示例项目，包括简单的 2 状态机、支付处理器、文本冒险游戏等。这些示例项目可以帮助用户了解 Orca 的使用方法。
Orca 需要 MCP 服务器环境变量来配置，包括 `ORCA_API_KEY`、`ORCA_PROVIDER`、`ORCA_BASE_URL` 和 `ORCA_MODEL` 等变量。
Orca 的工作流包括 TypeScript 包的构建和测试，以及 Go 包的构建和测试。
Orca 是一个有前景的 AI 工作流项目，值得关注
- 需要 orca-lang 解决具体问题的开发者与运营人员
- 先在测试环境跑通最小用例，再接入生产数据
- API key 直接提交到 git 仓库（请用 .env 并加入 .gitignore）
- 云端托管：可放在 Vercel / Railway / Fly.io 等 PaaS 平台
⚡ 核心功能
- 可视化 Agent 工作流编排，无需编写复杂代码
- 支持多步骤自动化任务链，实现全流程无人值守
- 与外部 API、数据库和第三方服务无缝集成
- 内置错误处理与自动重试机制，保障稳定运行
- 提供可复用的自动化模板，快速在同类场景部署
- 需要 orca-lang 解决具体问题的开发者与运营人员
- 先在测试环境跑通最小用例，再接入生产数据
- API key 直接提交到 git 仓库（请用 .env 并加入 .gitignore）
👥 适合人群
🎯 使用场景
- 自动化日常重复性工作，将精力集中于创造性任务
- 构建数据采集 → 处理 → 输出的完整自动化管线
- 实现跨平台、跨系统的数据流转和业务协同
⚖️ 优点与不足
- +大幅减少重复性人工操作
- +可视化流程，清晰直观
- +可扩展性强，支持复杂场景
- −未明确开源协议，商用场景需谨慎评估
- −初始配置和调试需投入一定时间
- −强依赖外部服务的稳定性
- −复杂场景需具备一定技术基础
该工具未明确声明开源协议，商业使用前请联系原作者确认授权范围，避免侵权风险。
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。
🔗 相关工具推荐
❓ 常见问题 FAQ
AI Skill Hub 点评：Orca 工作流 的核心功能完整，质量良好。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 | orca-lang |
| 原始描述 | 开源AI工作流：Orca (Orchestrated State Machine Language) — an LLM-native code generation targe。⭐14 · TypeScript |
| Topics | AI工作流TypeScript |
| GitHub | https://github.com/jascal/orca-lang |
| 语言 | TypeScript |
收录时间：2026-06-02 · 更新时间：2026-06-06 · License：未公布 · AI Skill Hub 不对第三方内容的准确性作法律背书。