# Unlimited OCR

> Local document parsing with baidu/Unlimited-OCR (MIT, 3B-total/500M-active MoE): markdown, LaTeX, tables and per-block layout bounding boxes, offline on Apple Silicon via MLX or an NVIDIA GPU via transformers. Ships the empirically-established working invocation (the upstream README's prompt decodes infinite garbage on MLX) and a figure segmenter that turns the model's refusal to describe charts into a pre-processing step.

## Facts
- Page: https://tashan.sh/capability/plugin-terrylica-cc-skills-unlimited-ocr
- tashan id: plugin:terrylica/cc-skills/unlimited-ocr
- Source: https://github.com/terrylica/cc-skills
- Type: plugin
- Category: cloud
- tashan score: 45.0 / 100
- Adoption: 7.0
- Upkeep: 99.0
- Freshness: 98.0
- Evidence coverage: 84% of the inputs this score can use
- Health: active
- Instruction depth: not yet graded
- License: MIT
- Official: no

## Install

```sh
/plugin marketplace add terrylica/cc-skills
/plugin install unlimited-ocr@cc-skills
```

## Security audit
Not scanned. We audit npm-published capabilities; this one has no npm package we can resolve, or has not reached the queue. This is not a clean bill of health.

---
Measured 2026-08-14 by tashan (https://tashan.sh) from public evidence. Scorer s5.
