cv

cv is a command line interface developed by cvfile.org for working with the .cv open file format, an open format designed for resume files. A .cv file is a valid PDF/A-3u document that additionally carries a clean Markdown copy, an optional HTML copy, and pre-computed BGE M3 semantic vectors as PDF Associated Files. This structure allows a single file to open visually in any PDF reader, parse cleanly in any applicant tracking system, embed on any webpage, and answer semantic queries from any retrieval pipeline. The software falls into the category of document tooling and developer utilities, with particular relevance to recruitment workflows, resume processing, and retrieval-augmented applications. The CLI ships the full reader and writer paths for the format: users can pack a PDF together with its text representations into a .cv file, extract a payload from an existing file, inspect XMP metadata, validate documents against the cv 1.0 specification, and run semantic search over the embedded vectors using the Hugging Face Inference API. Typical use cases include preparing resumes that remain both human-readable and machine-parseable, verifying that files conform to the specification before distribution, and performing semantic search across collections of resume files without re-computing embeddings. The current release is version 0.3.1, and one version is listed in the catalog. Beyond the CLI itself, the project ecosystem includes the full format specification, JavaScript and Python SDKs, a web component for embedding, and integrations with LangChain and LlamaIndex, all of which are maintained at https://github.com/cvfile/cv. Together these components support end-to-end workflows in which resumes function simultaneously as printable documents, structured data sources for applicant tracking systems, and vector-indexed records for semantic retrieval pipelines.

Tags: