PersonaFlow is an AI-powered research workflow system designed for Human-Computer Interaction (HCI) researchers and academics. It helps researchers generate, refine, and explore research questions through AI-driven personas, automated literature analysis, and iterative critique processes.
Check out our recent publication:
Liu, Y., Sharma, P., Oswal, M., Xia, H., & Huang, Y. (2025, July). PersonaFlow: Designing LLM-Simulated Expert Perspectives for Enhanced Research Ideation. Proceedings of the 2025 ACM Designing Interactive Systems Conference, 506-534.
- Research Question Generation: AI-powered generation of research questions from initial ideas
- Persona-Based Analysis: Generate AI researcher personas with different domain expertise
- Literature Discovery: Automatic literature search and analysis using Semantic Scholar API
- Iterative Critique: AI-driven critique and refinement of research ideas
- Research Outline Generation: Automated generation of research outlines and hypothetical abstracts
- Visual Node-Based Editor: Drag-and-drop interface for building research workflows
- Real-time Collaboration: Multi-user research discussion interface
- Interactive Tutorial: Guided onboarding for new users
- Export Capabilities: Export research outlines to Google Docs
- Support for multiple LLM providers (OpenAI, local models, custom endpoints)
- Semantic paper search and reranking
- Automated literature review generation
- Research scenario planning
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ Frontend │ │ Backend │ │ External │
│ (React) │◄───┤ (FastAPI) │◄───┤ APIs │
│ │ │ │ │ │
│ • Visual Editor │ │ • AI Chains │ │ • OpenAI API │
│ • User Auth │ │ • Literature │ │ • Semantic │
│ • Real-time UI │ │ • Agent System │ │ Scholar │
└─────────────────┘ └─────────────────┘ └─────────────────┘
│
┌─────────────────┐
│ Database │
│ (PostgreSQL │
│ + Supabase) │
└─────────────────┘
- Python 3.10+
- Node.js 20+
- PostgreSQL (or Supabase account)
- Redis (for caching)
- API Keys:
- OpenAI API key (or alternative LLM provider)
- Semantic Scholar API key
- Supabase credentials
-
Clone the repository
git clone https://github.com/your-username/personaflow.git cd personaflow -
Set up environment variables
# Copy example environment files cp backend/.env.example backend/.env.block cp backend/rds.env.example backend/rds.env # Edit the files with your API keys and database credentials nano backend/.env.block nano backend/rds.env
-
Start with Docker Compose
docker-compose up --build
-
Access the application
- Frontend: http://localhost:3000
- Backend API: http://localhost:8321
- Redis: localhost:6379
-
Set up Backend
cd backend/ # Create virtual environment python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate # Install dependencies pip install -r requirements.txt python -m spacy download en_core_web_sm # Set up environment variables cp .env.example .env.block cp rds.env.example rds.env # Edit these files with your credentials # Run the backend python main.py
-
Set up Frontend
cd frontend/rq-flow/ # Install dependencies npm install # Start development server npm run dev
-
Set up Database
- Create a PostgreSQL database
- Update
backend/rds.envwith your database credentials - The application will handle table creation automatically
# LLM Configuration
OPENAI_API_KEY=your_openai_api_key
OPENAI_API_BASE=https://api.openai.com/v1 # Optional: for custom endpoints
OPENAI_API_TYPE=openai # or 'azure' for Azure OpenAI
# Literature Search
S2_API_KEY=your_semantic_scholar_api_key
# Database (Supabase)
SUPABASE_URL=your_supabase_url
SUPABASE_SERVICE_KEY=your_supabase_service_key
SUPABASE_JWT_SECRET=your_jwt_secret
# Redis Configuration
REDIS_HOST=localhost
REDIS_PORT=6379
REDIS_PASSWORD= # Leave empty if no password
# Security
DECRYPT_KEY=your_32_character_encryption_key
# Optional: Alternative LLM Providers
XINFERENCE_API_URL=http://localhost:9997/v1
XINFERENCE_MODEL_ID=your_model_id
COHERE_API_URL=your_cohere_endpoint
COHERE_API_KEY=your_cohere_api_key# PostgreSQL/Supabase Database
RQGEN_DB_HOST=your_database_host
RQGEN_DB_PORT=5432
RQGEN_DB_USER=your_database_user
RQGEN_DB_PASS=your_database_password
RQGEN_DB_NAME=your_database_name- Semantic Scholar API: Get your key at https://www.semanticscholar.org/product/api#api-key
- OpenAI API: Register at https://platform.openai.com/
- Supabase: Create a project at https://supabase.com/
-
Start a New Research Project
- Click "Add RQ Node" to create a research question node
- Enter your initial research idea or question
-
Generate AI Personas
- Click "Next" to generate AI researcher personas
- Each persona represents a different research perspective
- Customize personas by editing their roles and backgrounds
-
Discover Literature
- Generate literature nodes to find relevant papers
- The system automatically searches Semantic Scholar
- Review and filter papers based on relevance
-
Generate Critiques
- Create critique nodes to analyze your research from different angles
- AI personas provide diverse critical perspectives
- Use critiques to identify gaps and opportunities
-
Refine Research Questions
- Generate new research question nodes based on critiques
- Iterate and refine your research focus
- Build complex research narratives
-
Create Research Outlines
- Generate detailed research outlines
- Create hypothetical abstracts
- Export to Google Docs for further development
- Group Discussions: Use the discussion panel for collaborative research
- Literature Analysis: Deep dive into paper relationships and citations
- Research Scenarios: Generate multiple research scenarios for comparison
- Progress Tracking: Monitor your research development over time
personaflow/
├── backend/ # FastAPI backend
│ ├── block_app/ # Core application logic
│ ├── autogpt/ # Agent system components
│ ├── db_utils/ # Database utilities
│ ├── routers/ # API route handlers
│ └── requirements.txt # Python dependencies
├── frontend/rq-flow/ # React frontend
│ ├── src/
│ │ ├── components/ # React components
│ │ ├── contexts/ # React contexts
│ │ ├── pages/ # Application pages
│ │ └── types/ # TypeScript types
│ └── package.json # Node.js dependencies
└── docker-compose.yaml # Docker configuration
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
# Backend tests
cd backend/
python -m pytest
# Frontend tests
cd frontend/rq-flow/
npm testOnce the backend is running, visit http://localhost:8321/docs for interactive API documentation powered by FastAPI's automatic OpenAPI generation.
POST /api/v1/block/persona_to_lit_query- Generate literature queries from personasPOST /api/v1/block/generate_literature_review- Generate literature reviewsPOST /api/v1/block/critique_to_rq- Generate research questions from critiquesPOST /api/v1/block/generate_hypothetical_abstract- Generate research abstracts
- Issues: Report bugs and request features via GitHub Issues
- Discussions: Join research discussions in GitHub Discussions
- Documentation: Additional documentation available in the
/docsfolder
This project is licensed under the MIT License - see the LICENSE file for details.
For questions about the research or system implementation, please reach out through GitHub Issues or contact the project maintainers:
Yiren Liu - yirenl2@illinois.edu
Note: This is a research prototype. While functional, it may require additional configuration and maintenance for production use.