Overview Version History Q & A Rating & Review
Offline Prompt Optimizer
Privacy-first VS Code extension for compressing prompts locally.
It helps you reduce token usage while preserving intent, with optional local AI assistance.
Features
Local prompt compression inside VS Code
Multiple compression modes: safe, balanced, aggressive
Token savings stats (before, after, saved)
Compression diff preview
Context menu actions on selected text
Optional Local AI setup for stronger compression
Works without cloud API keys for core rule-based compression
Requirements
Setup
Install dependencies:
npm install
Build extension bundle:
npm run build
Open this folder in VS Code and press F5 to launch the Extension Development Host.
Usage
Select prompt text in the editor (or use the full document).
Run one of:
Open Prompt Optimizer
Compress Selected Prompt
Preview Compression Diff
Review token savings and copy/apply compressed output.
Extension Commands
Command ID
Title
offline-prompt-optimizer.openPromptOptimizer
Open Prompt Optimizer
offline-prompt-optimizer.compressSelection
Compress Selected Prompt
offline-prompt-optimizer.previewCompressionDiff
Preview Compression Diff
offline-prompt-optimizer.copyCompressedPrompt
Copy Compressed Prompt
offline-prompt-optimizer.toggleCompressionMode
Toggle Compression Mode
offline-prompt-optimizer.setupLocalAI
Offline Prompt Optimizer: Setup Local AI
offline-prompt-optimizer.repairLocalAI
Offline Prompt Optimizer: Repair Local AI Setup
offline-prompt-optimizer.removeLocalAIAssets
Offline Prompt Optimizer: Remove Local AI Assets
offline-prompt-optimizer.showLocalAIStatus
Offline Prompt Optimizer: Show Local AI Status
Configuration
Settings namespace: offlinePromptOptimizer
offlinePromptOptimizer.defaultMode: safe | balanced | aggressive
offlinePromptOptimizer.defaultTokenizer: tokenizer for token counting
offlinePromptOptimizer.protectedPatterns: extra regex patterns protected from compression
offlinePromptOptimizer.localAI.enabled: enable AI-assisted local compression
offlinePromptOptimizer.localAI.runtime: llamafile | llama.cpp
offlinePromptOptimizer.localAI.modelProfile: nano | low-memory | balanced
Local AI Notes
nano profile uses SmolLM2 in-process (smallest setup).
Other profiles can download larger local runtime/model files.
Rule-based compression still works even if Local AI is not enabled.
Scripts
npm run build - Bundle extension to dist/extension.js
npm run watch - Bundle in watch mode
npm run compile - Type-check only (tsc --noEmit)
npm run test - Run tests
npm run package - Build .vsix package
Project Structure
new_extation/
src/ Extension source code
dist/ Bundled output
package.json Extension manifest
tsconfig.json TypeScript config
vitest.config.ts Test config
README.md Project documentation