🪶 OpenGrammar 2.0

Your privacy-first, open-source writing assistant.
The completely free, zero-compromise Grammarly alternative you can host yourself.

🌐 Official Website | 📖 Documentation Site

Writing clearly matters. But achieving perfect grammar shouldn't mean sacrificing your privacy or paying a premium monthly subscription to a tech giant.

OpenGrammar is a powerful intelligent writing assistant built to match the premium writing tools you already know, but with one massive difference: You own your data, and it runs completely for free.

🌟 Why OpenGrammar?

The Zero-Cost Foundation: Out of the box, our core engine runs 100% locally in your browser. With a massive 156,000-word offline dictionary, it instantly catches misspellings, passive voice, and weak phrasing without ever pinging a server.
Bring Your Own Brain (BYOK): Want deep, context-aware AI corrections? Just plug in your own API key (OpenAI, Groq, Together, OpenRouter, or local models via Ollama). You pay fractions of a cent exclusively for what you use.
Paranoid-Level Privacy: We don't have a database. We don't want your data. Your API key never leaves your browser storage. Text processed through the AI is immediately deleted via stateless edge functions.
Deploy It Yourself: You own the entire pipeline. Host our lightweight backend instantly via Docker, NPM, Cloudflare, Vercel, or Netlify.

🚀 Live Production Endpoints

Don't want to host the backend yourself? OpenGrammar comes with incredibly fast, globally-distributed public edge deployments that you can connect your browser extension to instantly:

☁️ Cloudflare Workers: https://cf.opengrammer.eu.cc/
△ Vercel Edge: https://vercel.opengrammer.eu.cc/
🟩 Netlify Functions: https://nl.opengrammer.eu.cc/

🛠️ Usage & Installation Environments

Our backend is natively multi-platform. Choose the environment that best fits your workflow.

📦 1. NPM / Mobile / Node Environments (Recommended for easy local use)

Run the OpenGrammar server on any machine with Node.js. You can effortlessly run this on an old laptop, a Raspberry Pi, or even an Android phone using Termux!

# Run instantly without installing
npx opengrammar-server

# Or install globally for persistent usage
npm i -g opengrammar-server

# Start the server on port 8787
opengrammar-server --port 8787

📱 Android (Termux) Setup:

Install Termux.
Run pkg update && pkg install nodejs.
Run npx opengrammar-server. Your phone is now a local grammar API!

🐳 2. Docker Setup

We maintain a highly optimized, multi-architecture Docker image. This is the recommended route for standard server environments or local homelabs.

# Pull the production image straight from Docker Hub
docker pull swdhinbiswas/opengrammar-backend:latest

# Run the container locally (exposes port 8787)
docker run -d -p 8787:8787 --name opengrammar swdhinbiswas/opengrammar-backend:latest

🧩 3. Browser Extension (For End Users)

Download the latest release from the opengrammar/extension/dist build.
Open Chrome/Brave/Edge and navigate to chrome://extensions/.
Enable Developer Mode.
Click Load unpacked and select the extension folder.
Click the OpenGrammar feather icon in your toolbar, input your custom AI API keys, and optionally link the Server URL to one of our Live Production Endpoints above!

🔌 API Reference & Environment Variables

If you are hosting OpenGrammar yourself via Docker or NPM, you have full access to its modular REST API endpoints.

Environment Variables (.env)

Create a .env file in your runtime directory to configure the AI routing logic.

Variable	Description
`PORT`	Server port (default: `8787`)
`GROQ_API_KEY`	Groq API key for ultra-fast AI grammar rewriting
`OPENAI_API_KEY`	OpenAI API key for premium AI models

Endpoints

Method	Endpoint	Description
`GET`	`/`	Status dashboard & Engine Version
`GET`	`/health`	Server health check
`POST`	`/analyze`	Advanced grammar analysis
`POST`	`/autocomplete`	AI Context-aware text completion
`GET`	`/providers`	Returns configured AI providers

Example Request:

curl -X POST http://localhost:8787/analyze \
  -H "Content-Type: application/json" \
  -d '{"text": "Me and him went to store yesterday."}'

✨ Application Features

The Dual-Engine Architecture: A blazing fast local (offline) RegEx + Dictionary matching engine, paired securely with a configurable LLM router framework for tone rewriting and smart context adjustments.
Deduplication Logic: The AI is fully synchronized with the local spellchecker. It never wastes expensive AI tokens re-checking typos already caught locally.
Smart Context Awareness: Automatically detects if you are writing a casual tweet, a technical document, or a formal business email, scaling its strictness appropriately.
Dynamic Writing Score: Evaluates your text dynamically (0-100) based on Correctness, Readability, Engagement, and Sentence Delivery.
Flawless Integrations: Securely injects into rich-text editors including Gmail, GoogleDocs, Notion, and Reddit.

🤝 Contributing

We need your help to unseat the monopolies and make OpenGrammar the definitive open-source writing assistant.

Whether you are a TypeScript developer wanting to optimize Vercel Edge latency, or a language nerd looking to build out our offline Regex GRAMMAR_RULES.md—we want you!

Read our CONTRIBUTING.md guide to get started locally.
Check our ROADMAP.md to see what features we are currently building.
Found a bug? Open an Issue. Have an idea? Start a thread in our Discussions tab.

📄 License

This project is proudly open-source under the Apache 2.0 License. See the LICENSE file for details. Let's build the future of writing together.