An AI-powered research assistant that transforms curiosity into structured research, complete with citations, task specifications, and actionable recommendations.

• Research Refinement: Transform raw queries into structured problem statements with Perplexity's Sonar models
• Task Specifications: Generate precise, actionable task specs with metrics and constraints
• SOTA Reviews: Access state-of-the-art research using academic search mode
• Technique Discovery: Find proven techniques from Kaggle, GitHub, and research papers
• Evidence-Based Recommendations: Get recommendations backed by citations and real-world examples
• Kaggle Integration: Explore datasets and competitions directly from the platform

• Perplexity API: Search API + Chat Completions (Sonar, Sonar-Pro, Sonar-Deep-Research, Sonar-Reasoning)
• Kaggle API: Dataset and competition exploration
• MongoDB: Research data persistence
• React + FastAPI: Modern full-stack architecture

supervisorctl restart backend
supervisorctl restart frontend

Backend Process: /var/log/supervisor/backend.err.log (or .out.log) Frontend Process: /var/log/supervisor/frontend.err.log (or .out.log)

• POST /api/research/refine - Refine research queries with Perplexity
• POST /api/research/task-spec - Generate task specifications
• POST /api/research/sota-review - Get SOTA reviews (academic mode)
• POST /api/research/techniques - Find techniques from Kaggle/GitHub
• POST /api/research/recommend - Get evidence-based recommendations
• GET /api/kaggle/datasets - Search Kaggle datasets
• GET /api/kaggle/competitions - List competitions