Smooth GPT

Fix ChatGPT lag, slow scrolling and typing in long conversations. Speeds up loading and reduces memory. Automatic — no setup needed.

As of June 2026, Smooth GPT has 20 users and a 5.00/5 rating from 3 reviews in the Productivity category.

Usersno change0%
20
20
Ratingno change0%
5.00
3 reviews
Reviewsno change0%
3
Version
1.0.0
Manifest V3

History

8 snapshots

Tracking since Apr 23, 2026.

21.28122.719999999999999Apr 23, 2026Jun 11, 2026
View as table
DateUsersRatingReviewsVersion
Apr 23, 20261.0.0
Apr 28, 20261.0.0
May 5, 202645.0021.0.0
May 16, 202655.0021.0.0
May 22, 2026105.0031.0.0
May 29, 2026125.0031.0.0
Jun 5, 2026105.0031.0.0
Jun 11, 2026155.0031.0.0
Now205.0031.0.0

Changelog

  • Apr 28, 2026
    description
    ChatGPT running slow? Smooth GPT kills the lag, stutter, and RAM drain in long conversations. Instant. Automatic.
    
    You're 80 messages deep into a brilliant conversation and ChatGPT has turned into a slideshow. Scrolling stutters. Typing lags. Your tab is eating 2 GB of RAM for breakfast.
    
    Smooth GPT fixes that. Install it, forget about it, enjoy a fast ChatGPT again.
    
    No setup. No buttons. No nonsense.
    
    ── WHAT'S ACTUALLY HAPPENING ──
    
    ChatGPT doesn't throw away old messages — it loads all 300 of them, renders every single one, and makes your browser suffer for it. Smooth GPT intercepts that before React gets anywhere near it.
    
    ▸ API trim — the big one
      Strips old messages from the conversation payload before they ever touch the DOM. A 280-message chat goes from ~27,000 DOM nodes down to ~4,000. Typing lag: gone. Scroll stutter: gone. You control how many messages to keep (default: 80).
    
    ▸ Code block deflation
      Got a chat full of code? Each syntax-highlighted block secretly contains hundreds of invisible <span> elements. Smooth GPT collapses off-screen code blocks to plain text and quietly rebuilds them as you scroll. DOM node count drops 60–90% in code-heavy conversations.
    
    ▸ Content-visibility + CSS containment
      Tells the browser to skip rendering off-screen messages entirely, and wraps each message so that typing in the input box can't cause a full-page reflow. The typing-lag fix that should've been in ChatGPT from day one.
    
    ▸ Aggressive pruning (optional, off by default)
      Full nuclear option. Physically removes off-screen messages from the DOM and replaces them with invisible placeholders. Flip it on if you're running a 500-message epic.
    
    ── PRIVACY — THE SHORT VERSION ──
    
    Zero data collected. Ever. Everything runs locally in your browser. No servers, no accounts, no tracking. Your conversations stay yours.
    
    ── PERMISSIONS — WHY WE NEED THEM ──
    
    • chatgpt.com & chat.openai.com — to actually do the work on ChatGPT pages.
    • scripting & tabs — to show live stats in the popup (DOM node count, messages trimmed, etc.).
    
    That's it. We don't touch any other site.
    
    ── OPEN SOURCE ──
    
    Built by Kozilabs. Say hi at kozilabs.lt
    ChatGPT running slow? Smooth GPT kills the lag, stutter, and RAM drain in long conversations. Instant. Automatic.
    
    You're 80 messages deep into a brilliant conversation and ChatGPT has turned into a slideshow. Scrolling stutters. Typing lags. Your tab is eating 2 GB of RAM for breakfast.
    
    Smooth GPT fixes that. Install it, forget about it, enjoy a fast ChatGPT again.
    
    No setup. No buttons. No nonsense.
    
    ── WHAT'S ACTUALLY HAPPENING ──
    
    ChatGPT doesn't throw away old messages — it loads all 300 of them, renders every single one, and makes your browser suffer for it. Smooth GPT intercepts that before React gets anywhere near it.
    
    ▸ API trim — the big one
      Strips old messages from the conversation payload before they ever touch the DOM. A 280-message chat goes from ~27,000 DOM nodes down to ~4,000. Typing lag: gone. Scroll stutter: gone. You control how many messages to keep (default: 80).
    
    ▸ Code block deflation
      Got a chat full of code? Each syntax-highlighted block secretly contains hundreds of invisible <span> elements. Smooth GPT collapses off-screen code blocks to plain text and quietly rebuilds them as you scroll. DOM node count drops 60–90% in code-heavy conversations.
    
    ▸ Content-visibility + CSS containment
      Tells the browser to skip rendering off-screen messages entirely, and wraps each message so that typing in the input box can't cause a full-page reflow. The typing-lag fix that should've been in ChatGPT from day one.
    
    ▸ Aggressive pruning (optional, off by default)
      Full nuclear option. Physically removes off-screen messages from the DOM and replaces them with invisible placeholders. Flip it on if you're running a 500-message epic.
    
    ── PRIVACY — THE SHORT VERSION ──
    
    Zero data collected. Ever. Everything runs locally in your browser. No servers, no accounts, no tracking. Your conversations stay yours.
    
    ── PERMISSIONS — WHY WE NEED THEM ──
    
    • chatgpt.com & chat.openai.com — to actually do the work on ChatGPT pages.
    • scripting & tabs — to show live stats in the popup (DOM node count, messages trimmed, etc.).
    
    That's it. We don't touch any other site.
    
    ── WHO DID THIS? ──
    
    Built by Kozilabs. Say hi at kozilabs.lt

Permissions & access

Permissions
scriptingtabs
Host access
https://chatgpt.com/*, https://chat.openai.com/*

Screenshots

Smooth GPT screenshot 1

About

ChatGPT running slow? Smooth GPT kills the lag, stutter, and RAM drain in long conversations. Instant. Automatic.

You're 80 messages deep into a brilliant conversation and ChatGPT has turned into a slideshow. Scrolling stutters. Typing lags. Your tab is eating 2 GB of RAM for breakfast.

Smooth GPT fixes that. Install it, forget about it, enjoy a fast ChatGPT again.

No setup. No buttons. No nonsense.

── WHAT'S ACTUALLY HAPPENING ──

ChatGPT doesn't throw away old messages — it loads all 300 of them, renders every single one, and makes your browser suffer for it. Smooth GPT intercepts that before React gets anywhere near it.

▸ API trim — the big one
  Strips old messages from the conversation payload before they ever touch the DOM. A 280-message chat goes from ~27,000 DOM nodes down to ~4,000. Typing lag: gone. Scroll stutter: gone. You control how many messages to keep (default: 80).

▸ Code block deflation
  Got a chat full of code? Each syntax-highlighted block secretly contains hundreds of invisible <span> elements. Smooth GPT collapses off-screen code blocks to plain text and quietly rebuilds them as you scroll. DOM node count drops 60–90% in code-heavy conversations.

▸ Content-visibility + CSS containment
  Tells the browser to skip rendering off-screen messages entirely, and wraps each message so that typing in the input box can't cause a full-page reflow. The typing-lag fix that should've been in ChatGPT from day one.

▸ Aggressive pruning (optional, off by default)
  Full nuclear option. Physically removes off-screen messages from the DOM and replaces them with invisible placeholders. Flip it on if you're running a 500-message epic.

── PRIVACY — THE SHORT VERSION ──

Zero data collected. Ever. Everything runs locally in your browser. No servers, no accounts, no tracking. Your conversations stay yours.

── PERMISSIONS — WHY WE NEED THEM ──

• chatgpt.com & chat.openai.com — to actually do the work on ChatGPT pages.
• scripting & tabs — to show live stats in the popup (DOM node count, messages trimmed, etc.).

That's it. We don't touch any other site.

── WHO DID THIS? ──

Built by Kozilabs. Say hi at kozilabs.lt

Technical

Version
1.0.0
Manifest
V3
Size
27.92KiB
Min Chrome
88
Languages
1
Featured
No

Metadata

ID
ofkacdopbdkjefpmicmjfdlbgjihdann
Developer ID
u50bd1425b730633e4a63ed8e3c117d4f
Developer Email
[email protected]
Created
Apr 22, 2026
Last Updated (Store)
Apr 24, 2026
Last Scraped
Jun 11, 2026
Website
Support URL

Data sourced from the Chrome Web Store · last verified Jun 11, 2026.