---
title: Arcee AI Trinity Large Code Review
emoji: 🔍
colorFrom: purple
colorTo: blue
sdk: docker
app_port: 7860
pinned: false
license: mit
---
# Arcee AI Trinity Large Code Review
AI-powered code review that streams real-time analysis of any public GitHub file using [Arcee AI's Trinity Large](https://huggingface.co/arcee-ai/trinity-large-preview) model via OpenRouter.
## How It Works
1. Paste a public GitHub file URL (e.g., `https://github.com/owner/repo/blob/main/file.py`)
2. Choose a review tone: **Gentle** (supportive mentor), **Standard** (Linus Torvalds-style), or **Brutal** (volcanic rage)
3. Get a streaming code review across 6 categories
## Review Categories
| Category | What It Covers |
|----------|---------------|
| **Summary** | High-level overview of the file's purpose and structure |
| **Code Quality** | Issues tagged as `[CRITICAL]`, `[HIGH]`, `[MEDIUM]`, or `[LOW]` severity |
| **Performance** | Algorithmic inefficiencies, memory leaks, unnecessary allocations |
| **Security** | Injection risks, authentication issues, data exposure |
| **Suggestions** | Concrete improvements with inline diff blocks |
| **Verdicts** | Opinions from Linus Torvalds, Donald Knuth, and Bjarne Stroustrup |
## Tech Stack
- **Backend**: Node.js + Express with Server-Sent Events (SSE)
- **Frontend**: Vanilla HTML/JS with tabbed interface, syntax highlighting, and diff rendering
- **Model**: `arcee-ai/trinity-large-preview:free` via [OpenRouter](https://openrouter.ai/)
- **Deployment**: Docker on Hugging Face Spaces
## Run Locally
```bash
git clone https://huggingface.co/spaces/juliensimon/trinity-code-reviewer
cd trinity-code-reviewer
echo "OPENROUTER_API_KEY=your_key_here" > .env
npm install
node server.js
```
Opens at [http://localhost:3000](http://localhost:3000).
## License
MIT