v0.1.1

Docs patch: the PyPI project page now renders the post-release README (PyPI install/quickstart forms — uvx paper-verify, pip install paper-verify). No code changes; 358 tests green.

The claude.ai web-chat skill zip is attached here as well.

v0.1.0

First public release on PyPI 🎉

Fact-check the citations in any document — extract every reference (URL / DOI / PMC / PMID / arXiv), fetch each source (metadata APIs → HTTP → Wayback), judge claim-vs-source with pluggable LLM judges, score on a transparent 100-point rubric, and report what a human should actually review.

Use it

uvx paper-verify yourdoc.md --level L2     # zero-install (uv)
pip install paper-verify                    # or install

No terminal? Copy-paste the web-chat prompt into any AI chat with browsing (한국어판), or upload paper-verify-webchat-skill.zip (attached below) to claude.ai as a Skill — extraction/scoring run as bundled code, fetching/judging use Claude's web tools, same rubric via --from-evidence.

Highlights

• Zero required dependencies (stdlib core); Anthropic/OpenAI/Gemini judges as optional extras
• --json agent surface, MCP server, harness profiles (claude-code / cursor / codex / gemini)
• --from-evidence / --extract-only: bring-your-own-fetch/judge, scored by the standard rubric
• 358 tests, feature-gate verified, PyPI trusted publishing + attestations

Full Changelog: https://github.com/nolainjin/paper-verify/commits/v0.1.0