# Changelog

## 0.1.0

- Add `pdf_parse` with Pi-authenticated model discovery, validated prompts, cost estimates, budgets, live progress, and durable artifacts.
- Add `pdf_search` over semantic parse blocks with a native PDF text-layer fallback.
- Add `pdf_screenshot` with durable PNG output and Pi image blocks.
- Add non-contiguous page selections and configurable render DPI.
- Add `/pdf-parse doctor` and compact tool rendering.