neuml/paperai

Implements RAG pipelines with configurable LLM inference to batch-process paper repositories and generate structured reports (Markdown, CSV, PDF annotations) based on semantic search queries and templated prompts. Built on txtai embeddings for vector indexing and supports both local GGUF models and remote LLMs for flexible deployment options. Configuration-driven architecture enables complex multi-column report generation with context-aware question answering across document collections.

1,747 stars and 987 monthly downloads. No commits in the last 6 months. Available on PyPI.