🔍
Multi-Source Search
Federated search across Semantic Scholar, OpenAlex, arXiv, and Crossref with automatic deduplication.
OmeletteAI-Powered Literature Management
Search · Deduplicate · OCR · Index · Chat — the complete research pipeline
Skip to content
Architecture
Omelette follows a modular pipeline architecture:
Keywords → Search → Dedup → Crawler → OCR → RAG → Writing
│ │ │ │ │ │ │
└─────────┴────────┴────────┴───────┴──────┴───────┘
LangGraph OrchestrationTech Stack
| Layer | Technology |
|---|---|
| Frontend | React 18 · TypeScript · Vite · TailwindCSS v4 · shadcn/ui |
| Backend | FastAPI · SQLAlchemy 2 (async) · Pydantic v2 · Python 3.12 |
| Database | SQLite + aiosqlite · Alembic |
| Vector Store | ChromaDB |
| RAG | LlamaIndex · BAAI/bge-m3 · bge-reranker-v2-m3 |
| LLM | LangChain (OpenAI · Anthropic · Aliyun · Volcengine · Ollama) |
| OCR | MinerU · pdfplumber · PaddleOCR |
| Pipeline | LangGraph with HITL interrupt/resume |
Testing
| Suite | Framework | Count |
|---|---|---|
| Backend | pytest-asyncio | 861 tests |
| Frontend | Vitest + Testing Library | 273 tests |
| E2E | Playwright | 39 tests |
| CI | GitHub Actions | All passing ✅ |