Smart Workflow

Smart Workflow is a powerful Obsidian plugin that enhances your knowledge management with AI-powered features and voice input.

中文文档

📢 Important Notice
The terminal feature has been extracted into a separate plugin: Obsidian Termy

Termy Features:

• Full terminal experience powered by xterm.js with Canvas/WebGL rendering
• Cross-platform support (Windows, macOS, Linux)
• Multiple shells: cmd, PowerShell, WSL, Git Bash, bash, zsh, custom shells
• Split panes (horizontal/vertical) and multiple sessions
• Search, font customization, theme support, background images
• Preset scripts for common workflows
• Rich keyboard shortcuts (Ctrl+O, Ctrl+Shift+R, Ctrl+F, etc.)

If you need terminal functionality, please install Termy instead.

✨ Features

🧠 AI Note Naming

• OpenAI-compatible API support (GPT, Claude, DeepSeek, Qwen, etc.)
• Multi-provider management with quick switching
• Custom prompt templates with variable injection
• Reasoning model support (auto-filters <think> tags)

🎤 Voice Input

• Push-to-talk dictation mode
• Multiple ASR engines: Alibaba Qwen, Doubao, SenseVoice
• Realtime streaming transcription
• LLM post-processing with custom presets

🌐 Translation

• Auto language detection
• Bidirectional translation (Chinese ↔ English)
• Selection toolbar integration

✍️ Writing Assistant

• Text polishing and refinement
• Streaming LLM responses
• Thinking process visualization

🚀 Installation

Manual Installation

1. Download main.js, manifest.json, styles.css from Releases
2. Place files in .obsidian/plugins/obsidian-smart-workflow/
3. Restart Obsidian and enable the plugin

Build from Source

git clone https://github.com/ZyphrZero/obsidian-smart-workflow.git
cd obsidian-smart-workflow

pnpm install
pnpm build
pnpm build:rust    # Build Rust server binary
pnpm install:dev   # Install to Obsidian

📖 Quick Start

Configure AI Provider

1. Go to Settings > AI Providers
2. Add a provider with endpoint and API key
3. Add models under the provider
4. Bind models to features (naming, translation, writing, etc.)

AI File Naming

• Command Palette: Ctrl/Cmd + P → "Generate AI File Name"
• Right-click Menu: Right-click file or editor

Voice Input

• Configure ASR credentials in settings
• Use hotkey to start/stop recording
• Transcription auto-inserts at cursor

⚙️ Configuration

Prompt Template Variables

{{content}}           - Note content (smart truncated)
{{currentFileName}}   - Current file name
{{#if currentFileName}}...{{/if}}  - Conditional block

Voice Settings

• ASR provider: Qwen / Doubao / SenseVoice
• Mode: Realtime (WebSocket) / HTTP
• Recording mode: Press-to-talk / Toggle
• LLM post-processing presets

🏗️ Architecture

┌─────────────────────────────────────────────────────────────┐
│                 Obsidian Plugin (TypeScript)                 │
├─────────────────────────────────────────────────────────────┤
│  Services                                                    │
│  ├── naming/       AI file naming                           │
│  ├── voice/        Voice input & ASR                        │
│  ├── translation/  Language detection & translation         │
│  ├── writing/      Writing assistant                        │
│  └── config/       Provider & model management              │
├─────────────────────────────────────────────────────────────┤
│  UI                                                          │
│  ├── settings/     Settings tabs                            │
│  ├── selection/    Selection toolbar                        │
│  └── voice/        Voice overlay                            │
└─────────────────────────────────────────────────────────────┘
                              │
                              │ WebSocket
                              ▼
┌─────────────────────────────────────────────────────────────┐
│              Smart Workflow Server (Rust)                    │
│  ├── voice/    Audio recording & ASR                        │
│  ├── llm/      LLM streaming                                │
│  └── utils/    Language detection                           │
└─────────────────────────────────────────────────────────────┘

🧩 FAQ

Q: Which AI providers are supported?
A: Any OpenAI-compatible API. Tested with OpenAI, Claude, DeepSeek, Qwen, GLM, etc.

Q: Voice input not working?
A: Check ASR credentials and ensure microphone permissions are granted.

🙏 Acknowledgements

• push-2-talk - Voice input architecture inspiration

---

Made with ❤️

⭐ Star this project if it helps you!