Keep ChatGPT fast by keeping only the last N messages in the DOM. Local-only, privacy-first browser extension that fixes UI lag in long conversations.

Long ChatGPT threads are brutal for the browser: the UI keeps every message in the DOM and the tab slowly turns into molasses — scroll becomes choppy, typing lags, devtools crawl.

LightSession fixes that by trimming old DOM nodes on the client side while keeping the actual conversation intact on OpenAI's side.

• Fixes UI lag in long chats
• Keeps model context intact (only the DOM is trimmed)
• 100% local – no servers, no analytics, no tracking

Built after too many coding sessions where a single ChatGPT tab would start eating CPU and turn Firefox into a slideshow.

• People who keep very long ChatGPT threads (100+ messages)
• Developers who use ChatGPT for debugging, code reviews, or long refactors
• Anyone whose ChatGPT tab becomes sluggish after a while in Firefox

Performance

• Automatic trimming – keeps only the last N messages visible (configurable range: 1–100 messages)
• DOM batching – node removals are batched within the ~16 ms budget for 60 fps scrolling
• Smart timing – waits for AI responses to fully finish streaming before trimming
• Ultra Lean Mode (Experimental) – aggressive optimizations: kills animations, applies CSS containment, dehighlights old code blocks

User experience

• Configurable – choose how many recent messages to keep (1–100)
• Status indicator – optional on-page pill showing trim statistics
• Reversible – refresh the page to restore the full conversation

Privacy

• Zero network requests – no data leaves your browser
• Local settings only – uses browser.storage.local for configuration
• No telemetry – no analytics, tracking, or usage reporting
• Domain-scoped – runs only on chat.openai.com and chatgpt.com

Install from AMO

After installation:

1. Open any ChatGPT conversation.
2. Click the LightSession icon in your Firefox toolbar.
3. Make sure the extension is enabled.
4. Adjust how many messages to keep if needed.

git clone https://github.com/11me/light-session.git
cd light-session

# Install dependencies
npm install

# Build the extension
npm run build

Then:

1. Open about:debugging#/runtime/this-firefox in Firefox.
2. Click Load Temporary Add-on.
3. Select extension/manifest.json.

1. Open a long ChatGPT conversation (or create one).
2. Click the LightSession toolbar icon.
3. Ensure Extension enabled is checked. Trimming will now happen automatically.
4. Use the slider to choose how many of the most recent messages to keep (1–100).

When you want to see the full history again:

• Click Refresh in the popup, or
• Reload the ChatGPT page.

• Extension enabled – master on/off toggle.
• Keep last N messages – how many messages remain visible in the DOM (1–100).
• Show status bar – display a floating pill with trim statistics.
• Ultra Lean Mode – aggressive performance mode for very long/laggy chats (experimental).
• Refresh – reloads the page to restore all messages.

• Navigate controls with Tab / Shift+Tab.
• Toggle checkboxes and buttons with Enter / Space.
• Adjust the slider with arrow keys.

No. LightSession only trims the DOM (what the browser renders), not the data stored by OpenAI.

• The conversation on OpenAI's servers remains intact.
• Reloading the page (or using Refresh in the popup) restores the full history.

Yes:

• No external network requests are made by the extension.
• No analytics, tracking, or telemetry.
• Settings are stored locally in browser.storage.local.

LightSession uses a multi-tier selector strategy and conservative fallbacks, but a major UI redesign may temporarily break trimming. In that case:

• The extension will simply stop trimming (fail-safe).
• Your conversations will continue to work as usual.
• An update will be released to restore trimming behavior.

LightSession uses a non-destructive trimming pipeline:

1. Detection – finds ChatGPT message nodes with a multi-tier selector system (data attributes → test IDs → structural + heuristic fallback).
2. Classification – labels nodes as user / assistant / system / tool messages.
3. Calculation – determines which messages to keep based on your settings.
4. Batching – removes excess nodes in small chunks using requestIdleCallback to stay within the frame budget.
5. Markers – optionally leaves comment markers in the DOM for debugging.

Trimming only affects what the browser renders. The conversation itself remains on OpenAI's side and is fully recoverable by reloading the page.

• Node.js >= 24.10.0 (see .node-version)
• npm >= 10
• Firefox >= 115

# Install dependencies
npm install

# Build once
npm run build

# Watch + rebuild on changes
npm run watch

# Run tests
npm run test

# Lint
npm run lint

# Format
npm run format

# Run in Firefox Developer Edition
npm run dev

# Run in Firefox (stable)
npm run dev:stable

# Package for distribution
npm run package

# Clean build artifacts
npm run clean

extension/
├── src/
│   ├── content/        # Content scripts (run on ChatGPT pages)
│   ├── background/     # Background script (settings management)
│   ├── popup/          # Popup UI (HTML/CSS/JS)
│   └── shared/         # Shared types, constants, utilities
├── dist/               # Compiled output (TypeScript → JavaScript)
├── icons/              # Extension icons
└── manifest.json       # Firefox extension manifest

• State machine for the trimmer: IDLE ↔ OBSERVING (simplified two-state design)
• Debounced MutationObserver (~75ms) to batch DOM changes
• Idle callback (requestIdleCallback) for non-blocking node removal
• Fail-safe thresholds (e.g. minimum message count) to avoid over-trimming

• Browser: Firefox >= 115 (Manifest V3)
• OS: Windows, macOS, Linux
• ChatGPT: Optimized for the current UI (2025), resilient to small layout changes

This repository contains the Firefox implementation. A separate Chrome version is available on the Chrome Web Store.

Pull requests are welcome. For larger changes or new features, please open an issue first to discuss the approach.

MIT License - see LICENSE for details.

• Issues: GitHub Issues

If you find this extension useful, you can support ongoing development:

Disclaimer: This is an unofficial extension not affiliated with OpenAI.