AI-powered paper writing assistant for computer science researchers.
PaperKiller is an integrated tool that helps researchers with four key aspects of paper writing:
- Ideation: Brainstorm research ideas, analyze innovation potential, and generate research questions
- Experiment: Generate experimental code and execute in sandboxed environments
- Diagram: Create framework diagrams (AI-generated) and visualize experimental results (matplotlib)
- Writing: AI-assisted academic writing with quality checking and citation generation
- Backend: Python 3.11+, FastAPI
- Frontend: Vue 3, TypeScript, Naive UI
- AI: Anthropic Claude API (primary), OpenAI DALL-E 3 (image generation)
- Visualization: matplotlib, seaborn
- Execution: Docker containerized code execution
paperkiller/
├── backend/ # Python FastAPI backend
│ ├── app/
│ │ ├── api/ # API route handlers
│ │ ├── core/ # Core business logic modules
│ │ ├── models/ # Data models
│ │ ├── services/ # External service wrappers (Claude, OpenAI)
│ │ └── utils/ # Utility functions
│ ├── tests/ # Backend tests
│ └── requirements.txt
├── frontend/ # Vue 3 frontend
│ ├── src/
│ │ ├── components/ # Reusable components
│ │ ├── views/ # Page views (Home, Ideation, Experiment, Diagram, Writing)
│ │ ├── stores/ # Pinia state stores
│ │ └── types/ # TypeScript definitions
│ └── package.json
└── docker-compose.yml # Development environment
- AI-powered brainstorming with multiple frameworks (SWOT, 5W1H, First Principles, Design Thinking)
- Innovation analysis with novelty assessment
- Research question generation
- Methodology suggestions
- Code generation for experiments (Python, R, Julia, JavaScript)
- Sandboxed code execution using Docker
- Result analysis and interpretation
- Experiment suggestions based on research questions
- AI-optimized prompt generation for diagram creation
- DALL-E 3 integration for framework diagrams
- matplotlib/seaborn chart generation for experimental results
- Chart type suggestions based on data and research questions
- AI content generation for all paper sections
- Text refinement with specific goals
- Writing quality checking
- Paper structure suggestions
- Citation generation in multiple styles (APA, IEEE, Chicago, MLA, BibTeX)
- Python 3.11+
- Node.js 20+
- Docker (optional, for code execution sandbox)
- Anthropic API key (required)
- OpenAI API key (optional, for DALL-E image generation)
-
Clone the repository
git clone <repository-url> cd PaperKiller
-
Backend setup
cd backend cp .env.example .env # Edit .env and add your ANTHROPIC_API_KEY # Optionally add OPENAI_API_KEY for DALL-E features pip install -r requirements.txt uvicorn app.main:app --reload
-
Frontend setup (in a new terminal)
cd frontend cp .env.example .env npm install npm run dev -
Access the application
- Frontend: http://localhost:5173
- Backend API: http://localhost:8000
- API Documentation: http://localhost:8000/docs
# Start all services
docker-compose up --build
# Start specific service
docker-compose up backend
docker-compose up frontendOnce the backend is running, visit http://localhost:8000/docs for interactive API documentation.
Ideation (/api/ideation)
POST /brainstorm- Generate research ideasPOST /analyze-innovation- Analyze innovation potentialPOST /generate-questions- Generate research questionsPOST /suggest-methodology- Suggest research methodology
Experiment (/api/experiment)
POST /generate-code- Generate experimental codePOST /execute- Execute code in sandboxPOST /analyze-results- Analyze experimental resultsPOST /suggest-experiments- Suggest experimental approaches
Diagram (/api/diagram)
POST /generate-prompt- Generate optimized image promptPOST /generate-image- Generate AI image (DALL-E 3)POST /plot-chart- Generate matplotlib chartPOST /suggest-chart- Suggest appropriate chart types
Writing (/api/writing)
POST /assist- Generate or refine academic textPOST /refine- Refine text with specific goalsPOST /check-quality- Check writing qualityPOST /suggest-structure- Suggest paper structurePOST /generate-citation- Generate formatted citationsPOST /compile- Compile LaTeX to PDF
| Variable | Description | Default |
|---|---|---|
ANTHROPIC_API_KEY |
Your Anthropic API key | Required |
OPENAI_API_KEY |
OpenAI API key for DALL-E | Optional |
DEBUG |
Enable debug mode | false |
DATABASE_URL |
Database connection string | SQLite |
DOCKER_ENABLED |
Enable Docker for code execution | true |
FRONTEND_URL |
CORS allowed origin | http://localhost:5173 |
| Variable | Description | Default |
|---|---|---|
VITE_API_BASE_URL |
Backend API URL | http://localhost:8000/api |
VITE_APP_NAME |
Application name | PaperKiller |
Backend:
cd backend
pytestFrontend:
cd frontend
npm run testBackend (uses Black and Ruff):
cd backend
black app/
ruff check app/Frontend:
cd frontend
npm run lintPhase 1-5 are complete. The application now has:
- ✅ Full Claude API integration
- ✅ All four modules (Ideation, Experiment, Diagram, Writing) with real AI functionality
- ✅ Complete frontend with all pages
- ✅ Backend API with comprehensive endpoints
- ✅ Code generation and (simulated) execution
- ✅ matplotlib chart generation
- ✅ DALL-E 3 integration (when OPENAI_API_KEY is configured)
- ✅ Writing assistance with multiple features
- Code execution uses simulation when Docker is not available
- LaTeX compilation requires a full TeX installation (included in Docker)
- Some advanced features may need additional configuration
- Phase 6: Full integration and optimization
- User authentication and project persistence
- Collaborative editing features
- More LaTeX templates
- Export to various journal formats
- Enhanced error handling and user feedback
Contributions are welcome! Please feel free to submit a Pull Request.
MIT License - see LICENSE file for details.