ray.so Code Snippet Image Generator

Generate beautiful code snippet images using ray.so and save them locally.

Requirements

• The user MUST provide the code snippet, either directly or by pointing to a file/selection in context
• MUST ask the user for ALL styling parameters before generating, presenting ALL available options
• MUST use agent-browser for screenshot capture (check availability first)

Workflow

Step 1: Verify agent-browser Availability

Before proceeding, verify that agent-browser is available:

which agent-browser

If agent-browser is not found in the PATH, inform the user that this skill requires agent-browser and cannot proceed without it.

Step 2: Fetch Available Options

Fetch the current themes and languages from ray.so's GitHub repository using curl:

# Fetch and parse available themes
curl -s "https://raw.githubusercontent.com/raycast/ray-so/main/app/(navigation)/(code)/store/themes.ts" | grep -oE 'id:\s*"[^"]+"' | sed 's/id:\s*"//;s/"//' | sort -u

# Fetch and parse available languages
curl -s "https://raw.githubusercontent.com/raycast/ray-so/main/app/(navigation)/(code)/util/languages.ts" | grep -oE '^[[:space:]]*"?[a-zA-Z0-9+#-]+"?\s*:\s*\{' | sed 's/[[:space:]]*"//g;s/".*//;s/:.*//' | sort -u

Step 3: Ask User for ALL Parameters

MUST use AskUserQuestion to ask for EVERY parameter, presenting ALL available options. Ask for parameters in this order:

3.1 Theme Selection

Present ALL available themes. In the question, list every theme fetched from step 2. Example:

Question: "Which theme would you like?"
Description: "Available themes: [list ALL themes from curl output]"
Options (pick 4 popular ones for quick select):
- breeze (default, purple gradient)
- midnight (cyan-blue)
- vercel (minimalist dark)
- sunset (warm orange)
Note: User can select "Other" to type any theme from the full list

3.2 Language Selection

Infer the language when possible. Skip this question if:

• The user explicitly specified a language
• The code comes from a file with a clear extension (e.g., .py → python, .js → javascript, .ts → typescript, .rs → rust, .go → go, etc.)
• The syntax is unmistakably identifiable (e.g., def/import → python, func/package → go, fn/let mut → rust)

Only ask this question if the language cannot be confidently inferred:

Question: "Which language for syntax highlighting?"
Description: "Available languages: [list ALL languages from curl output]"
Options:
- auto (auto-detect)
- javascript
- python
- typescript
Note: User can select "Other" to type any language from the full list

3.3 Dark/Light Mode

Question: "Dark or light mode?"
Options:
- Dark mode (default)
- Light mode

3.4 Background

Question: "Show the gradient background?"
Options:
- Yes, show background (default)
- No, transparent/minimal background

3.5 Padding

Question: "How much padding around the code?"
Options:
- 16 (compact)
- 32 (small)
- 64 (medium, default)
- 128 (large)

3.6 Line Numbers

Question: "Show line numbers?"
Options:
- No (default)
- Yes

3.7 Title

Question: "Add a title above the code? (e.g., filename)"
Options:
- No title (default)
- Yes, add title
If yes, ask for the title text.

Note: Do NOT ask about output path/filename. Save to the current working directory with a sensible filename (e.g., rayso-snippet.png, or based on the title if provided like fibonacci.png). Only use a different path if the user explicitly specifies one in their original request.

Step 4: Build the ray.so URL

CRITICAL: ALL parameters must be in the URL hash (after #), NOT in the query string.

Build the URL using shell commands:

# 1. Base64 encode the code
CODE_BASE64=$(echo -n 'YOUR_CODE_HERE' | base64)

# 2. URL encode the base64 string
CODE_ENCODED=$(python3 -c "import urllib.parse; print(urllib.parse.quote('$CODE_BASE64'))")

# 3. Build the URL with ALL parameters in the hash
# Format: https://ray.so/#param1=value1&param2=value2&code=ENCODED_CODE
# Do NOT include width parameter - let ray.so auto-size to fit content
URL="https://ray.so/#theme=THEME&padding=PADDING&background=BACKGROUND&darkMode=DARKMODE&language=LANGUAGE&code=${CODE_ENCODED}"

# Add optional parameters if needed:
# If lineNumbers: add "&lineNumbers=true" before &code=
# If title: add "&title=URL_ENCODED_TITLE" before &code=

URL Hash Parameters:

Note on width: Do NOT include the width parameter unless you specifically need a fixed width. Without it, ray.so auto-sizes the frame to fit the code content, avoiding unnecessary empty space.

Example URL construction:

# For code: for i in range(23):\n    print(i)
# Theme: midnight, Padding: 64, Dark mode: true, Background: true, Language: python, Title: test.py

CODE='for i in range(23):
    print(i)'
CODE_BASE64=$(echo -n "$CODE" | base64)
CODE_ENCODED=$(python3 -c "import urllib.parse; print(urllib.parse.quote('$CODE_BASE64'))")
TITLE_ENCODED=$(python3 -c "import urllib.parse; print(urllib.parse.quote('test.py'))")
URL="https://ray.so/#theme=midnight&padding=64&background=true&darkMode=true&language=python&title=${TITLE_ENCODED}&code=${CODE_ENCODED}"
echo "$URL"

Step 5: Capture High-Quality Image with agent-browser

MUST use agent-browser (verified in Step 1). This approach uses the html-to-image library (same as ray.so's internal export) with high pixelRatio for crisp, sharp text rendering.

IMPORTANT: Always use a unique session name with --session to avoid stale session issues.

# Generate unique session name
SESSION="rayso-$(date +%s)"

# 1. Set viewport
agent-browser --session $SESSION set viewport 1400 900

# 2. Open the URL
agent-browser --session $SESSION open "$URL"

# 3. Wait for the page to fully render
agent-browser --session $SESSION wait --load networkidle
agent-browser --session $SESSION wait 3000

# 4. Load html-to-image library (same library ray.so uses internally)
agent-browser --session $SESSION eval 'new Promise((r,e)=>{const s=document.createElement("script");s.src="https://cdn.jsdelivr.net/npm/html-to-image@1.11.11/dist/html-to-image.js";s.onload=r;s.onerror=e;document.head.appendChild(s)})'

# 5. Capture at 4x resolution using html-to-image (produces crisp text)
agent-browser --session $SESSION eval 'htmlToImage.toPng(document.querySelector("#frame > div"),{pixelRatio:4,skipAutoScale:true})' > /tmp/rayso-dataurl-$SESSION.txt

# 6. Close the browser
agent-browser --session $SESSION close

# 7. Convert data URL to PNG file
DATAURL=$(cat /tmp/rayso-dataurl-$SESSION.txt | tr -d '"' | tr -d '\n')
echo "$DATAURL" | sed 's/data:image\/png;base64,//' | base64 -d > /path/to/output.png

# 8. Clean up temp file
rm /tmp/rayso-dataurl-$SESSION.txt

Critical notes:

• Uses html-to-image library which is what ray.so uses for its own export feature
• pixelRatio: 4 produces high-DPI images with crisp, sharp text (4x native resolution)
• The data URL is captured directly from the library, not from a screenshot
• No ImageMagick required - pure browser-based rendering at high resolution
• Output is correctly sized with no extra whitespace

Step 6: Confirm Output and STOP

Report the saved file location to the user. The task is complete - do not perform any additional checks, explorations, or verifications after the screenshot is saved.

Complete Example

User: "Create a code snippet image of this Python function"

def fibonacci(n):
    if n <= 1:
        return n
    return fibonacci(n-1) + fibonacci(n-2)

1. Check which agent-browser - confirmed available

1. Fetch themes and languages:

curl -s "https://raw.githubusercontent.com/raycast/ray-so/main/app/(navigation)/(code)/store/themes.ts" | grep -oE 'id:\s*"[^"]+"' | sed 's/id:\s*"//;s/"//' | sort -u

1. Ask user for parameters via AskUserQuestion:

• Theme: user selects "midnight"
• Language: inferred as python from def syntax - not asked
• Dark mode: user selects "Dark mode"
• Background: user selects "Yes"
• Padding: user selects "64"
• Line numbers: user selects "No"
• Title: user selects "No title"

1. Build URL (all params in hash, no width for auto-sizing):

CODE='def fibonacci(n):
    if n <= 1:
        return n
    return fibonacci(n-1) + fibonacci(n-2)'
CODE_BASE64=$(echo -n "$CODE" | base64)
CODE_ENCODED=$(python3 -c "import urllib.parse; print(urllib.parse.quote('$CODE_BASE64'))")
URL="https://ray.so/#theme=midnight&padding=64&background=true&darkMode=true&language=python&code=${CODE_ENCODED}"

1. Capture high-quality image:

SESSION="rayso-$(date +%s)"

agent-browser --session $SESSION set viewport 1400 900
agent-browser --session $SESSION open "$URL"
agent-browser --session $SESSION wait --load networkidle
agent-browser --session $SESSION wait 3000

# Load html-to-image library
agent-browser --session $SESSION eval 'new Promise((r,e)=>{const s=document.createElement("script");s.src="https://cdn.jsdelivr.net/npm/html-to-image@1.11.11/dist/html-to-image.js";s.onload=r;s.onerror=e;document.head.appendChild(s)})'

# Capture at 4x resolution
agent-browser --session $SESSION eval 'htmlToImage.toPng(document.querySelector("#frame > div"),{pixelRatio:4,skipAutoScale:true})' > /tmp/rayso-dataurl-$SESSION.txt
agent-browser --session $SESSION close

# Save as PNG
DATAURL=$(cat /tmp/rayso-dataurl-$SESSION.txt | tr -d '"' | tr -d '\n')
echo "$DATAURL" | sed 's/data:image\/png;base64,//' | base64 -d > ./fibonacci.png
rm /tmp/rayso-dataurl-$SESSION.txt

1. Report: "Saved code snippet image to ./fibonacci.png"

Image Resolution and Quality

This skill uses the html-to-image library with pixelRatio: 4 to produce high-quality images with crisp, sharp text. This is the same rendering approach that ray.so uses for its built-in export feature.

Output quality:

• Default: 4x native resolution (frame auto-sizes to content, then rendered at 4x)
• Text is rendered at high DPI, not upscaled from low resolution
• Gradient backgrounds and all CSS styling are preserved
• No unnecessary empty space (frame auto-sizes to fit code)

Adjusting resolution:

• For smaller files: Change pixelRatio:4 to pixelRatio:2 in the eval command
• For maximum quality: Use pixelRatio:6 (same as ray.so's "6x" export option)

Forcing a specific width:

• Only add &width=NUMBER to the URL if you need a fixed width (e.g., for consistent sizing across multiple images)

Troubleshooting

• If agent-browser is not available: Inform the user and do not proceed
• If curl fails to fetch themes/languages, use these common defaults:
• Themes: breeze, midnight, candy, crimson, falcon, meadow, raindrop, sunset, vercel, supabase, tailwind
• Languages: auto, javascript, typescript, python, rust, go, java, ruby, swift, kotlin, css, html, json, yaml, bash
• If parameters aren't applied: Ensure ALL parameters are in the URL hash (after #), not the query string
• If title isn't showing: The title parameter must be in the hash: #title=filename.py&code=...
• If html-to-image fails to load: Check network connectivity; the library loads from jsdelivr CDN
• If capture returns empty: The frame selector #frame > div may have changed; inspect the page structure
• For very long code snippets, ray.so may truncate; consider splitting into multiple images
• If the page doesn't load properly, increase the wait time (try 4000ms or more)
• If you get a blank page: Use a fresh unique session name with --session flag
• If data URL is malformed: Ensure quotes and newlines are stripped before base64 decoding