Converge the SDK, Gradio, remote CLI, Doclib and Router on Tasks with hash-based uploads, per-file result URLs, task listing and optional model output. Preserve Files and the synchronous file_parse wrapper while removing standalone Uploads/Jobs and local-path HTTP inputs.
Migrate contract, lifecycle, transport and real-document regression coverage; update documentation and HTTP examples. Validation: 1452 passed, 3 conditional skips, plus targeted final PDF checks.
BREAKING CHANGE: clients and workers must use Tasks; standalone Uploads/Jobs endpoints and local-source/inline HTTP options have been removed.
Accept form manifests and per-file uploads with resumable task state and optional source hashes. Publish result URLs only after final ZIP artifacts are ready, and align Router handling with the same protocol.
Remove synchronous parsing and task result routes, drop fixed local upload limits, and update the HTTP client, documentation, and regression coverage.
- Remove MLX references from user-facing docs (README, docs/zh, docs/en,
skills/mineru/SKILL.md); macOS VLM docs now state only the llama.cpp default
- Restructure READMEs: README_en.md becomes the main README.md with the skill
frontmatter, README_zh.md becomes README_zh-CN.md, and the Agent Guide is
embedded as a collapsible section in both
- Link DocVortex in feature lists; update Acknowledgments (add DocVortex and
metafile-render, remove UniMERNet, fast-langdetect, pdftext)
- Remove the MinerU-Diffusion citation entry
- Drop the development-era mineru-kit gradio wording from the 4.0 migration table
- Remove the solved Linux missing-text FAQ entry
- Drop Noto font packages and fc-cache from the 4.0 Dockerfiles (keep libgl1)
myhloli