xint

Project overview

Vexor creates reusable semantic indexes over files and source code so users can locate relevant material with natural-language queries. The same engine is available through a CLI, a Python API, and a built-in MCP server, with configurable embedding and reranking providers, hybrid keyword-plus-semantic ranking, and fully local model options. For agents such as Claude Code or Codex, the included skill and MCP server expose the same core search capability, so a workflow generally needs only one of those integrations.

Project overview

paper-fetch is a dependency-free Python CLI and agent skill that resolves a DOI—or a title first matched through Crossref and Semantic Scholar—into a PDF using a multi-source fallback chain. It supports batch files or stdin, NDJSON streaming, idempotent retries, consistent filenames, and checks for SSRF, valid PDF bytes, and size limits. The final fallbacks can include publisher access and Sci-Hub mirrors, which are enabled by default as a last resort, so users should review source selection and their access rights.