Markdown & Component Reference

Type Scale

Display / H1 — 3rem / 800

Page Heading — 2.25rem / 700

Section Heading — 1.2em / 700 / Caps

H3 — 1.5rem / 700 / Accent

H4 — 1.25rem / 600

Body — 1.0625rem / 400 — Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris.

Muted — Same size, muted color — Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.

Mono — 1rem — code blocks, terminal output, file paths

Headings in Context

Heading 2 — The Architecture Challenge

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.

Heading 2 — Component Selection

Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

Heading 3 — Document Processing Pipeline

Handles ingestion, chunking, and embedding generation for the knowledge base. The pipeline processes documents through multiple stages to extract structured information.

Heading 4 — Chunking Strategy

Sed ut perspiciatis unde omnis iste natus error sit voluptatem accusantium doloremque laudantium, totam rem aperiam.

Heading 5 — Implementation Notes

Nemo enim ipsam voluptatem quia voluptas sit aspernatur aut odit aut fugit.

Callouts

📝 Note

Lorem ipsum dolor sit amet, consectetur adipiscing elit. This is a general informational callout that draws attention to supplementary details the reader might find useful.

💡 Tip

Ut enim ad minim veniam, quis nostrud exercitation. Tips highlight best practices or shortcuts. For example, you can press Ctrl+K to open the command palette in most editors.

⚠️ Warning

Duis aute irure dolor in reprehenderit. Warnings flag potential pitfalls. Double-check your filing deadlines — missing them can reset the entire process.

🔴 Danger

Excepteur sint occaecat cupidatat non proident. Running this command will permanently delete all local data. This action cannot be undone.

Dense Data Table

Model	Params	MMLU	HumanEval	GSM8K	Format	License	Apple Silicon
Qwen2.5-7B-Instruct	7.6B	74.2	84.1	83.8	MLX 4-bit	Apache 2.0	M1+
Llama-3.1-8B-Instruct	8.0B	73.0	72.6	84.4	MLX 4-bit	Llama 3.1	M1+
Gemma-2-9B-IT	9.2B	71.3	54.3	76.7	MLX 4-bit	Gemma	M2+
Phi-3.5-mini-instruct	3.8B	69.0	62.8	86.2	MLX 4-bit	MIT	M1+
Mistral-7B-Instruct-v0.3	7.2B	62.5	38.4	52.2	MLX 4-bit	Apache 2.0	M1+
Yi-1.5-9B-Chat	8.8B	69.5	46.3	68.2	MLX 4-bit	Apache 2.0	M2+

Simple Table

Item	Count	Status
Detection pipeline	7 stages	Complete
YOLO model	18 MB	Trained
Unit tests	134	Passing
Benchmarks	3 datasets	Evaluated

Data Cards

uitag v0.4.2

1.2s

Avg detection time per screen

94.3% accuracy

MIT license

vLLM-MLX

Models validated on M-series

6 VLMs

3 chip gens

Interagent

Agent entities coordinated

44 projects

3 runtimes

Bento Grid (3x1)

🔍

Detect

Apple Vision identifies text regions. VLM handles icons and non-text elements selectively.

🏷️

Label

Every element gets a bounding box, semantic label, and pixel coordinates. Machine-readable in under a second.

🤖

Act

Downstream agents use structured output to click, type, and navigate — no pixel guessing, no prompt hacking.

Accessible Form

Subscribe to updates

Prose Elements

Italic text and bold text and bold italic text.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.

Ut enim ad minim veniam, quis nostrud exercitation.

Lists

Lorem ipsum dolor sit amet
Consectetur adipiscing elit
- Sed do eiusmod tempor
- Incididunt ut labore
Ut enim ad minim veniam

First step in the process
Second step follows naturally
1. Sub-step with detail
2. Another sub-step
Third step completes the sequence

Detection pipeline complete
YOLO model trained
ScreenSpot-Pro full evaluation
Documentation finalized

Code

Inline code: pip install uitag or python -m pytest.

// siteConfig.ts
export const SITE = {
  website: "https://swaylenhayes.com/",
  author: "Swaylen Hayes",
  desc: "AI tooling, Apple ecosystem, automation.",
  title: "~/swaylen",
  lightAndDarkMode: true,
  postsPerPage: 8,
};

# inference.py
def run_detection(image_path: str, model: str = "qwen2-vl-7b") -> dict:
    """Run hybrid UI detection on a screenshot."""
    text_regions = apple_vision.detect_text(image_path)
    icon_regions = vlm.detect_icons(image_path, model=model)
    return merge_regions(text_regions, icon_regions)

Links and References

Here is a link to a page and an external link.

Here’s a footnote, ¹ and here’s another one. ²

Inline Elements

Ctrl + Shift + P

Do or do not. There is no try.

H₂O and E=mc²

~~Strikethrough text~~ that is no longer relevant.

VLM and SoM

Horizontal Rules

Content above the rule.

Content below the rule.

Another section.

Lorem ipsum dolor sit amet, first footnote. ↩
Consectetur adipiscing elit, second footnote. ↩

Hybrid UI Detection: Why We Split Vision and Intelligence

Benchmarking UI Detection on ScreenSpot-Pro