down-craft
Node.js package to simplify the process of converting documents (PDF, DOCX, PPTX, and XLSX) into Markdown format. It uses tesseract.js, mammoth, pdf.js, and turndown to convert documents to Markdown format. For PDFs, it also provides an option to use vLLMs (Vision Large Language Models) for advanced OCR capabilities (using the OpenAI API).
A Node.js package that converts documents — PDF, DOCX, PPTX, and XLSX — into Markdown. It uses tesseract.js, mammoth, pdf.js, and turndown for conversion, and for PDFs it offers a choice of standard text extraction, Tesseract OCR, or vLLM-based OCR via the OpenAI API for higher-fidelity results. The main downCraft(fileBuffer, fileType?, options?) function accepts a document buffer and optional file type (auto-detected when omitted) and returns the Markdown string. The vLLM converter extracts embedded images, renders PDF pages to high-quality images, runs vLLM OCR, and cleans up temporary files, with LLM parameters configurable directly or via environment variables.
Installation
npm install down-craftRelated \\
-
bedrock-proxy-endpoint
Spin up your own custom OpenAI API server endpoint for easy AWS Bedrock LLM text inference (using standard baseUrl, and apiKey params)
-
bedrock-wrapper
Bedrock Wrapper is an npm package that simplifies the integration of existing OpenAI-compatible API objects with AWS Bedrock's serverless inference LLMs.…
-
InferIQ
LLM Evaluation framework that leverages LLMs to Evaluate other LLMs