Tox World: Block the Rage Bait with a Giant Sun

The Problem

Some days you're sharp. Twitter rage bait bounces off. Reddit wars look silly. YouTube comment sections are comedy.

Other days you're vulnerable. Someone's angry reply ruins your morning. Doom scroll turns into actual doom. Bot army rage seeps in.

Current solution: Willpower. Self-control. "Just don't read the comments."

Actual solution: Stop reading them. Make software read them for you.

The Concept

Browser extension that scans every page. Detects toxic content. Replaces it with a giant sun SVG.

Not hidden (you know something's there, curiosity kills you).

Replaced. Sun appears. Message: "Blocked: Someone's having a bad day."

Visual peace. Zero poison.

Some Days You Want the Chaos

That's the key insight.

Monday: Feeling strong. Let the bots rage. Entertainment.

Wednesday: Burnt out. Need protection. Enable the sun.

Toggle button. Toolbar icon. Click on, click off. Your choice. Your mood. Your boundary.

Not a permanent filter. A variable shield based on how resilient you feel that day.

How It Works (Technical)

Browser Extension (Manifest V3):

1. Content script runs on every page automatically
2. Scans all text nodes (comments, posts, replies, headlines)
3. Sentiment analysis - catches toxic language
4. Replaces matched content with sun SVG
5. User toggle in toolbar (on/off, whitelist sites)
6. Settings panel for sensitivity levels

The Hybrid Approach (Pragmatic):

Don't call an API for every comment. Expensive. Slow. Unnecessary.

Local keyword filter catches 90% of rage:

• Obvious insults ("idiot," "stupid," "stfu," "kys")
• Rage patterns (all caps, excessive punctuation, slurs)
• Bot signatures (repeated phrases, copy-paste attacks)

Fast. Free. Private. Runs in browser. No API calls.

Optional AI check for borderline cases:

• User clicks "Check this one" on false positives
• Extension sends to Claude/OpenAI API
• AI determines: toxic or just passionate?
• User trains their own threshold

Best of both: Speed + intelligence. Privacy + accuracy.

The Sun Replacement

Not just hide the toxic content. Replace it.

<div class="tox-world-block">
  <svg width="120" height="120" viewBox="0 0 120 120">
    <circle cx="60" cy="60" r="30" fill="#FFD700"/>
    <g stroke="#FFD700" stroke-width="3">
      <line x1="60" y1="10" x2="60" y2="25"/>
      <line x1="60" y1="95" x2="60" y2="110"/>
      <line x1="10" y1="60" x2="25" y2="60"/>
      <line x1="95" y1="60" x2="110" y2="60"/>
      <line x1="25" y1="25" x2="35" y2="35"/>
      <line x1="85" y1="85" x2="95" y2="95"/>
      <line x1="85" y1="25" x2="95" y2="35"/>
      <line x1="25" y1="85" x2="35" y2="95"/>
    </g>
  </svg>
  <p>Blocked: Someone's having a bad day</p>
</div>

Big. Friendly. Obvious. Calm.

You see the sun. You know what was there. You don't absorb the poison.

Prompting Strategy for Building It

This is a "specific fucking prompting" situation. AI can build browser extensions fast if you're exact.

Vague (AI struggles):

"Make an extension that blocks bad comments"

Exact (AI crushes):

"Manifest V3 Chrome extension. Content script scans text nodes on page load. Regex matches toxic keywords (list: idiot, stupid, stfu, kys, + all caps patterns). Replace matched DOM elements with inline SVG sun (gold #FFD700, 120px). Add toolbar toggle icon (on/off state). Settings page: whitelist domains, sensitivity slider (strict/normal/off). Use chrome.storage.sync for settings persistence."

Technical specification. Testable. AI executes perfectly.

Prompting the Sentiment Filter

Vague:

"Detect toxic language"

Exact:

"Regex patterns: (1) Insults - case-insensitive match: idiot|stupid|moron|dumb|loser. (2) Rage - all caps 10+ chars, 3+ exclamation marks. (3) Slurs - blocklist array. (4) Bot signatures - exact string repeats 3+ times. Return boolean: isToxic true/false."

AI knows exactly what to build.

Prompting the UI

Vague:

"Make it look nice"

Exact:

"Sun SVG: 120px circle, gold #FFD700 fill, 8 rays as lines radiating. Below: text 'Blocked: Someone's having a bad day' in 14px gray sans-serif. Whole component: centered flex column, 200px height, fade-in 0.3s ease. Background: light yellow #FFFEF0 with 1px dotted border #DDD."

Pixel-perfect spec. AI executes.

The Whitelist

You don't want to block your friends. Just the rage bots.

Settings panel:

• Whitelist specific domains (trust your Substack writers)
• Whitelist specific users (Twitter friends, Reddit communities)
• Blacklist extra keywords (custom triggers for your life)

Your filter. Your rules. Your peace.

Local vs. API: The Decision

Local keyword matching:

• Pros: Instant, free, private, works offline
• Cons: Dumb, misses nuance, false positives

AI API (Claude, OpenAI):

• Pros: Understands sarcasm, context, subtlety
• Cons: Costs money, API latency, rate limits

Hybrid solution:

1. Local filter runs on everything (catches 90%)
2. User can right-click blocked content: "Check with AI"
3. Extension sends that one comment to API
4. AI verdict: "Actually toxic" or "False positive - just passionate"
5. User trains their personal threshold

Best of both. Free for most cases. Smart when you need it.

Why This Works

You're not blocking the internet. You're blocking your own vulnerability.

Strong day? Turn it off. Let the chaos entertain you.

Burnt out? Turn it on. Let the sun protect you.

Variable shield. Not censorship. Not avoidance. Just: tactical boundary based on current resilience.

Same way you don't answer your phone when exhausted. Same way you don't go to loud parties when overstimulated.

Digital boundaries that match your actual state.

The Absurdity

We need browser extensions to avoid other humans' bad days.

That's where we are.

Internet gave everyone a megaphone. Algorithms reward rage. Bots amplify anger. Humans having normal bad days look like apocalyptic meltdowns.

You're not weak for needing a filter. You're pragmatic.

Peace isn't a character flaw.

Build It Yourself

Step 1: Prompt Claude or ChatGPT with exact specs (see "Prompting Strategy" above)

Step 2: Test locally (Chrome: Load unpacked extension in chrome://extensions)

Step 3: Add your own keyword list (your triggers, your life)

Step 4: Publish to Chrome Web Store (or keep it private, whatever)

Step 5: Toggle it on when you need peace

Step 6: Toggle it off when you want chaos

Step 7: Realize you keep it on more than you expected

The Meta Joke

This article could be toxic depending on your mood.

Cynical take on internet culture? Rage-adjacent commentary? Passive aggressive at bot boys?

Make your own extension and block me.

That's how you know it works.

---

Build Tox World. Protect your peace. Watch the sun rise on a quieter internet.

Some days you want the fight. Other days you just want the sun.

Code gives you the choice.