A system-wide AI writing assistant. Use your own LLM API keys - or run grammar checks locally - no subscription, no vendor lock-in, full control.
🌐 中文文档
If you already pay for LLM API access, you can use it as a writing assistant; no need for additional subscriptions to services like DeepL or Grammarly. Today's flash models can polish and translate text at a fraction of a cent per request.
WritingTools is the most popular open-source alternative. It follows an Apple Intelligence-inspired pop-up menu workflow (select text -> pick action -> get result). PhrAIse takes a different approach: one hotkey opens a persistent floating window in which all the styles, rewrites and grammar fixes are visible at once, with no need for extra clicks between edits.
Neither is objectively better. They're built for different writing habits. This is the one that fits mine.
Select text anywhere, hit a hotkey (Default: Ctrl+C+C), and PhrAIse rewrites or translates it using your configured engine.
PhrAIse has two interchangeable engines for the Optimize function:
Simple and Fast
Harper is a local, offline grammar engine. No API key, no network request, no data leaves your machine. It catches spelling mistakes, repeated words, missing/extra spaces, wrong articles (a/an), unclosed quotes, apostrophe issues, and more.
When Harper is assigned to Optimize, PhrAIse shows a single corrected version and a list of detected grammar issues. You can toggle individual fixes on or off before replacing the original text.
Demo Opmization (Harper mode)
More Powerful
When you assign a configured LLM model to Optimize or Translate, PhrAIse uses your LLM provider:
- Optimize: three rewrite versions per request with grammar checking, in any style you define (Concise, Formal, Natural, or custom)
- Translate: translation with auto-detection, one-click replace
- Custom instructions: free-form AI edits with extensible style presets
Demo Optimization (LLM mode)
Demo Translation (LLM mode)
- System integration: global hotkeys, UIA-based text grab (clipboard-safe), system tray, app blacklist
- Configuration: OpenAI-compatible API, dual model slots, per-function model assignment, custom CSS theming
The built-in provider list is fetched from models.dev on startup and merged with a curated local fallback. This means new OpenAI-compatible endpoints appear automatically; if the network is unavailable, PhrAIse falls back to the local list.
⚠️ Coding Plan WarningMost providers define what their coding plans / subscriptions may be used for according to their own documentation. "Coding plans" are generally intended for interactive programming agents (such as OpenCode, Cursor, Codex, etc.), NOT for writing assistants like PhrAIse. Using a coding plan in PhrAIse may violate the provider's terms of service and could result in your account being suspended or banned.
Please check Coding Plan Status whether PhrAIse is allowed to use their coding plans.
💵 If you're looking for free LLM API, the following repos may help you:
Settings are in %APPDATA%/PhrAIse/settings.json. All configurable through the Settings panel.
PhrAIse supports built-in color themes and custom CSS overrides. See the Custom Theming Guide for details on selecting themes, writing override CSS, and adding new palettes.
| Hotkey | Action |
|---|---|
Ctrl+C+C (double-tap C) |
Trigger optimize/translate on selected text |
Ctrl+Shift+B |
Toggle floating ball visibility |
Esc |
Close floating window |
Hotkeys are configurable in Settings.
If you just want to run PhrAIse without installing Python, download the latest release from the GitHub Releases page and run PhrAIse.exe directly.
👉 On first launch, right-click the tray icon or floating ball -> Settings -> configure at least one model with your API key, or set Optimize to Harper to use the local grammar engine.
- Windows 10 (21H2+) / 11
- Python 3.11+
- For LLM mode: API key from any OpenAI-compatible LLM provider
- For Harper local mode: nothing extra; the Harper binary is bundled
pip install -r phraise/requirements.txt
python run.py👉 On first launch, right-click the tray icon or floating ball -> Settings -> configure at least one model with your API key, or set Optimize to Harper to use the local grammar engine.
Package into a standalone .exe with PyInstaller:
python build.py
# Output: dist/PhrAIse.exeReleases will be published on GitHub with downloadable .exe files. The download badge at the top of this README will show total downloads once the release workflow is active.
Phraise/
├── build.py # PyInstaller build script
├── phraise/
│ ├── main.py # Application entry point & lifecycle
│ ├── config.py # JSON-based configuration management
│ ├── i18n.py # Internationalization
│ ├── theme.py # Themes & CSS config
│ ├── prompts.py # LLM prompt templates
│ ├── llm_client.py # OpenAI-compatible API client
│ ├── harper_client.py # Local Harper grammar engine client
│ ├── harper_lsp_manager.py # Harper LSP subprocess manager
│ ├── settings_panel.py # Settings dialog (models, styles, triggers, appearance)
│ ├── floating_ball.py # Draggable always-on-top orb
│ ├── floating_window.py # Main window (optimize / translate)
│ ├── hotkeys.py # Global hotkey listener
│ ├── text_grabber.py # UIA text extraction & replacement
│ ├── detector.py # Active window detection
│ ├── dispatch.py # Thread-safe main-thread dispatcher
│ ├── error_log.py # Error logging
│ └── assets/ # Icons and resources
└── README.md
PhrAIse is free and open-source under the GNU GPLv3. No premium tiers, no subscriptions, no limitations.
If you find it useful, the best way to support the project is:
- 🐞 Report bugs or unexpected behavior
- 💡 Suggest features or improvements
- 🔧 Submit pull requests
- 📢 Share it with someone who might find it helpful
Built for my own workflow. Shared in case it fits yours too.