home / skills / partme-ai / full-stack-skills / ascii-image-to-ascii

ascii-image-to-ascii skill

Q: Does the skill fetch or upload images for me?

No. The script requires a local imagePath; it does not fetch or upload images on your behalf.

Q: What inputs should I provide?

Provide imagePath, optional width (default 100), charset (default " .:-=+*#%@"), mode (readable or detail), background hint, and colorMode (none or ansi256).

Q: How do I improve noisy output?

Crop tightly to the subject, increase contrast, simplify the background, and reduce width or use a smaller charset for the readable variant.

safe

/skills/ascii-image-to-ascii

This skill converts a local image into ASCII art with readable and detail variants, offering width, charset, and optional ANSI for previews.

npx playbooks add skill partme-ai/full-stack-skills --skill ascii-image-to-ascii

Review the files below or copy the command above to add this skill to your agents.

Files (4)

SKILL.md

2.6 KB

---
name: ascii-image-to-ascii
description: Convert an image into ASCII art (readable + detail variants, width/charset controls, optional ANSI), for terminal previews and plain-text “image substitutes”.
license: Complete terms in LICENSE.txt
dependencies:
  - python>=3.8
  - pillow
---


## When to use this skill
**CRITICAL TRIGGER RULE**
- Use this skill ONLY when the user explicitly mentions the exact skill name: `ascii-image-to-ascii`.

**Trigger phrases include:**
- "ascii-image-to-ascii"
- "use ascii-image-to-ascii"
- "用 ascii-image-to-ascii 把图片转字符画"
- "使用 ascii-image-to-ascii 生成可读优先/细节优先两版"

## Boundary
- Default output is ASCII-only; ANSI coloring is optional.
- The bundled script is for local conversion/verification only. Do not fetch/upload images on the user’s behalf.
- Always give pre-processing advice (crop subject, raise contrast, simplify background) before generating the final ASCII art.

## How to use this skill
### Inputs
- imagePath (local path, required)
- width (default 100; common: 80/100/120)
- charset (default ` .:-=+*#%@`, from light to dark)
- mode (readable | detail; if outputting both, this can be ignored)
- background (light | dark | unknown, default unknown)
- colorMode (none | ansi256, default none)

### Outputs (required)
- asciiReadable: readable-first (less noise, clearer silhouette)
- asciiDetail: detail-first (more levels, richer texture)
- paramsGuide: recommended width/charset + pre-processing tips
- pitfalls: 3-5 common failure modes with fixes

### Steps
1. Pre-processing advice (required):
   - Crop the subject, remove irrelevant background
   - Increase contrast to reduce gradient noise
   - Target width usually 80–120 columns
2. Charset + mapping direction:
   - Space is the lightest, `@` is the darkest (or invert consistently)
3. Generate two variants:
   - readable: fewer levels (less noise)
   - detail: more levels (more texture)
4. Optional ANSI:
   - Use color only as a hint; keep the silhouette readable
   - Always provide a no-color fallback

## Script
- `scripts/image_to_ascii.py`
  - Requires Python + Pillow (`pip install pillow`)
  - Supports: readable/detail variants, optional `--ansi256`

## Examples
- `examples/readable-vs-detail.md`

## Quality checklist
1. Lines are consistent and `<= width`
2. The readable variant must preserve the main silhouette
3. No trailing spaces (copy/paste safe)

## Keywords
**English:** ascii-image-to-ascii, image to ascii, ascii art, terminal preview, charset, grayscale, dithering, ansi
**中文:** ascii-image-to-ascii, 图片转字符画, ASCII 字符画, 终端预览, 字符集, 灰度映射, 降噪, ANSI 彩色

Overview

This skill converts a local image file into two ASCII-art variants: a readable-first silhouette-preserving version and a detail-first texture-rich version. It supports width and charset controls and optional ANSI256 coloring while always providing a no-color fallback. The tool is intended for local use and includes preprocessing guidance to improve results.

How this skill works

The script loads a local image, applies optional preprocessing (crop, contrast), resizes to the target column width while preserving aspect ratio, and maps image luminance to a configurable character set. It produces two outputs: asciiReadable (fewer levels to keep silhouettes clear) and asciiDetail (more levels for texture). ANSI256 coloring can be added as a hint but a plain-ASCII fallback is always generated.

When to use it

When you explicitly invoke ascii-image-to-ascii by name
To create terminal previews or plain-text substitutes for images
When you need both a silhouette-focused and a texture-focused ASCII version
When distributing ASCII art that must be copy/paste safe and terminal-friendly
To test visual results locally without uploading images

Best practices

Preprocess: crop to the subject, increase contrast, and simplify the background before conversion
Choose width 80–120 columns; common defaults are 80, 100, 120
Select charset from light-to-dark (default: " .:-=+*#%@"); use fewer levels for readable variant and more for detail variant
Verify lines are <= width and remove trailing spaces for safe copy/paste
If using ANSI coloring, treat color as a hint; always keep a plain-ASCII fallback

Example use cases

Generate a readable preview of a product photo for terminal display
Create a detail-rich ASCII texture version for artistic text UIs
Produce copy/paste-friendly ASCII placeholders for text-only environments
Compare silhouette vs. texture output when refining preprocessing steps
Embed ASCII art into README or terminal dashboards with optional ANSI color

FAQ

Does the skill fetch or upload images for me?

No. The script requires a local imagePath; it does not fetch or upload images on your behalf.

What inputs should I provide?

Provide imagePath, optional width (default 100), charset (default " .:-=+*#%@"), mode (readable or detail), background hint, and colorMode (none or ansi256).

How do I improve noisy output?

Crop tightly to the subject, increase contrast, simplify the background, and reduce width or use a smaller charset for the readable variant.