shbshahriar/parallel-writing-assessment-engine repository overview

A containerised FastAPI application that evaluates writing quality across 8 dimensions simultaneously using a LangGraph parallel pipeline and Claude (claude-haiku-4-5). Dimensions scored: grammar, clarity, coherence, depth, structure, vocabulary, tone, and readability. Also performs argument structure analysis and generates a structured improvement report with strengths, weaknesses, a revision plan, and a band prediction.

Serve the UI and API from a single port: docker run -p 8000:8000 --env-file .env writing-assessment-engine

Requires ANTHROPIC_API_KEY (and optionally OPENAI_API_KEY, GOOGLE_API_KEY, LANGCHAIN_API_KEY) passed in at runtime via --env-file. See .env.example.