# Evidence

One page per claim, with its method and limits. The repeatable hardware measurements use one Comet and one mini PC, taken on 2026-09-29 and 2026-09-30. The NAS recovery is a separate operator account and post-mortem, without an independent action log.
Raw test records are described on each page. A short video from the later Comet boot capture is published on the home page; the full raw frame set is not published.

Read in this order for the strongest proof first:

1. **Real recovery**: [nas-recovery](nas-recovery.md).
2. **Platforms**: [platforms](platforms.md).
3. **OCR**: [ocr-role](ocr-role.md), then [bios-ocr](bios-ocr.md) and [vlm-vs-ocr](vlm-vs-ocr.md).

| Page | Claim in one line |
|---|---|
| [nas-recovery.md](nas-recovery.md) | A Comet with local OCR was used to diagnose and fix a real boot loop on a headless NAS |
| [platforms.md](platforms.md) | Linux tests recorded; Windows appears in an operator account without a repeatable suite |
| [ocr-role.md](ocr-role.md) | Why local OCR mattered in the recovery, using numbers from the pages below |
| [mouse-accuracy.md](mouse-accuracy.md) | The target's kernel sees the mouse within 1/32767; X within 1 px |
| [typing.md](typing.md) | 25/25 exact strings, fast and slow |
| [bios-ocr.md](bios-ocr.md) | On a real BIOS, RapidOCR 0.98, Tesseract 0.94 |
| [vlm-vs-ocr.md](vlm-vs-ocr.md) | Strict phrases of 32: local VLM 31, Tesseract 22, RapidOCR 16 |
| [boot-menu-catch.md](boot-menu-catch.md) | Setup reached on a real AMI BIOS by tapping a key through POST |
| [comet-quirks.md](comet-quirks.md) | What the Comet does differently from upstream PiKVM |

Common limits: one measured Comet setup, one test target, one operator. Nothing here is a vendor-wide success rate.