glance-cli 0.8.0 → 0.8.2

package/CHANGELOG.md

@@ -5,6 +5,21 @@ All notable changes to glance-cli will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.8.2] - 2026-01-02
+
+### Changed
+- Added Bun as preferred installation method in documentation
+- Updated CLI imports for better compatibility
+
+### Fixed
+- Removed development files from distribution
+
+## [0.8.1] - 2026-01-02
+
+### Changed
+- Streamlined README for better npm presentation
+- Minor documentation improvements
+
 ## [0.8.0] - 2026-01-02
 
 ### Added

package/README.md

@@ -1,17 +1,13 @@
 # glance-cli
 
-**AI-powered web reader for your terminal** – Fetch any webpage, extract clean content, summarize with AI (local or cloud), or ask questions about it.
+**AI-powered web reader for your terminal** – Fetch any webpage, extract clean content, and get instant AI summaries.
 
-- **100% FREE by default**: Auto-detects and uses Ollama (local AI) - no API keys needed!
-- **Local-first**: Run offline with Ollama models (e.g., Llama 3, Mistral, Gemma)
-- **Optional cloud support**: OpenAI (GPT series) and Google Gemini for premium quality
-- **Smart fallback**: Automatically uses best available service (prefers free)
-- **Voice output**: Listen to summaries with free system TTS or premium ElevenLabs
-- **Production-ready**: Expert prompt engineering, comprehensive error handling, intelligent caching
-- **Privacy-focused**: Your data stays local with Ollama - no cloud required
-- **Built with Bun**: Lightning-fast startup, TypeScript, battle-tested
+- **100% FREE by default** – Uses local Ollama (no API keys needed!)
+- **Privacy-first** – Your data stays on your machine
+- **Lightning fast** – Built with Bun and TypeScript
+- **Production-ready** – Smart caching, error handling, and retry logic
 
-`glance` turns any webpage into terminal-friendly insights — no browser needed.
+Turn any webpage into terminal-friendly insights — no browser needed.
 
 [![npm version](https://badge.fury.io/js/glance-cli.svg)](https://www.npmjs.com/package/glance-cli)
 [![Downloads](https://img.shields.io/npm/dm/glance-cli.svg)](https://www.npmjs.com/package/glance-cli)
@@ -19,504 +15,256 @@
 
 ---
 
-## ✨ Highlights
-
-### 🎯 **Expert Prompt Engineering**
-Advanced prompt templates engineered for superior AI results:
-- **Structured outputs** with quality constraints
-- **Anti-hallucination** measures (only facts from source)
-- **Task-specific optimization** (TL;DR, key points, ELI5, Q&A)
-- **10x better results** than generic prompts
-
-### 🛡️ **Production-Grade Reliability**
-Built for real-world use:
-- **Automatic retries** with exponential backoff
-- **Timeout protection** (no infinite hangs)
-- **Comprehensive error handling** with actionable hints
-- **Graceful degradation** on failures
-- **Memory leak prevention** (guaranteed resource cleanup)
-
-### 💾 **Intelligent Caching**
-Smart caching for speed and cost savings:
-- **TTL-based expiration** (24h default, configurable)
-- **Automatic compression** (70% space savings)
-- **LRU eviction** (100MB max, configurable)
-- **Cache statistics** (hit rate, size, performance metrics)
-- **Tag-based organization** for easy management
-
-### 📊 **Rich Content Extraction**
-Advanced content detection and extraction:
-- **Intelligent scoring algorithm** (picks best content, not first)
-- **Structure preservation** (headings, lists, paragraphs)
-- **Comprehensive metadata** (20+ fields including author, dates, reading time)
-- **Table extraction** with structure preservation
-- **Code block detection** with language identification
-- **Link categorization** (internal/external/anchor)
-
-### 🎨 **Multiple Output Formats**
-Export in any format you need:
-- **Terminal** – Colorful, emoji-rich display
-- **Markdown** – Perfect for documentation
-- **JSON** – Structured data for APIs
-- **HTML** – Styled, embeddable webpages
-- **Plain Text** – Script-friendly, pipeable
-
-### 🎙️ **Voice/TTS Output**
-Listen to summaries instead of reading:
-- **ElevenLabs integration** – Premium, natural-sounding voices
-- **Local TTS fallback** – Works offline (macOS Say, Windows SAPI, Linux espeak)
-- **Audio export** – Save summaries as MP3 files
-- **Multiple voices** – Choose from various voice profiles
-- **No API key required** – Falls back to system TTS automatically
-
----
-
-## 📦 Installation
+## ⚡ Quick Start
 
-### Via npm (global)
 ```bash
+# Install with Bun (preferred - faster!)
+bun install -g glance-cli
+
+# Or with npm
 npm install -g glance-cli
-```
 
-### Via Bun (recommended – fastest)
-```bash
-bun add -g glance-cli
+# Use immediately (works without setup!)
+glance https://example.com
 ```
 
-### Standalone Binary (no Node/Bun required)
-Download pre-compiled binaries from [Releases](https://github.com/jkenley/glance-cli/releases).
-
-**macOS/Linux:**
+**For 100% free local AI:**
 ```bash
-curl -L https://github.com/jkenley/glance-cli/releases/latest/download/glance-$(uname -s)-$(uname -m) -o /usr/local/bin/glance
-chmod +x /usr/local/bin/glance
-```
+# One-time setup (~4GB download)
+curl -fsSL https://ollama.com/install.sh | sh
+ollama pull llama3
 
-**Windows:**
-Download the `.exe` from [Releases](https://github.com/jkenley/glance-cli/releases) and add to PATH.
+# Now completely free forever
+glance https://nextjs.org/docs --tldr
+```
 
 ---
 
-## 🆓 Free Setup (No API Keys!)
+## 🎯 Why Glance?
 
-Want to use glance completely free? Just install Ollama:
+| Feature | Glance + Ollama | Traditional AI Tools |
+|---------|-----------------|---------------------|
+| **Cost** | $0 forever | $20-100/month |
+| **Privacy** | 100% local | Cloud-based |
+| **Speed** | 2-3 seconds | 10-30 seconds |
+| **Setup** | 2 minutes | Instant (with $$) |
 
-```bash
-# 1. Install Ollama (one-time setup)
-curl -fsSL https://ollama.com/install.sh | sh  # macOS/Linux
-# Or download from: https://ollama.com/download
-
-# 2. Download a model (one-time, ~4GB)
-ollama pull llama3
-
-# 3. That's it! Glance will auto-detect and use it
-glance https://example.com
-```
-
-**Why Ollama?**
-- ✅ **100% FREE** - No API costs ever
-- ✅ **Private** - Runs entirely on your machine
-- ✅ **No internet required** - Works offline after model download
-- ✅ **Fast** - Optimized for your hardware (CPU/GPU)
+Perfect for developers who want fast, private, and free AI summarization in their terminal.
 
 ---
 
-## ⚡ Quick Start
+## 📖 Basic Usage
 
-### Basic Usage (Works Without Any API Keys!)
+### Summary Modes
 ```bash
 # Quick summary
-glance https://example.com
+glance https://nextjs.org/docs
 
-# Ask a specific question
-glance https://react.dev --ask "What are React Server Components?"
+# One-sentence TL;DR
+glance https://techcrunch.com/article --tldr
+
+# Bullet point key insights
+glance https://blog.example.com --key-points
 
-# Get key points with emojis
-glance https://blog.example.com --key-points --emoji
+# Simple explanation
+glance https://complex-topic.com --eli5
 
-# TL;DR in one sentence
-glance https://news.ycombinator.com --tldr
+# Ask specific questions
+glance https://nextjs.org/docs --ask "What's the App Router?"
+```
 
-# Listen to summary (text-to-speech)
-glance https://article.com --read
+### AI Models
 
-# Save audio summary
-glance https://podcast.com --key-points --audio-output summary.mp3
+**Free Local AI (Recommended):**
+```bash
+glance https://example.com --model llama3
+# Popular models: llama3, mistral, phi3, gemma2
+# List available: glance --list-models
 ```
 
-### Local Models (Free & Private)
+**Optional Cloud AI:**
 ```bash
-# Use Ollama with Llama 3
-glance https://example.com --model llama3 --stream
+# OpenAI (requires API key)
+export OPENAI_API_KEY=sk-...
+glance https://example.com --model gpt-4o-mini
 
-# Available models: llama3, mistral, phi3, gemma2, codellama, etc.
-# List installed models:
-glance --list-models
+# Google Gemini (requires API key)
+export GEMINI_API_KEY=...
+glance https://example.com --model gemini-2.0-flash-exp
 ```
 
-### Check Available Services
+**Smart Detection:**
 ```bash
-# See what services you have available
+# See what services you have
 glance --check-services
 
-# Force free-only mode (never use paid APIs)
+# Force free-only
 glance https://example.com --free-only
 
-# Prefer quality (use paid APIs if available)
+# Prefer quality (use paid if available)
 glance https://example.com --prefer-quality
 ```
 
-### Advanced Features
+### Output Options
 ```bash
-# Export to Markdown with metadata
-glance https://example.com --export article.md --include-metadata
+# Export to Markdown
+glance https://nextjs.org/docs --export notes.md
 
-# Capture screenshot + summary
-glance https://example.com --screenshot page.png
+# Export to JSON
+glance https://api-docs.com --json -o data.json
 
-# Full render for JavaScript-heavy sites
-glance https://react-app.com --full-render
+# Save as audio (text-to-speech)
+glance https://article.com --audio-output summary.mp3
 
-# Multiple languages supported
-glance https://lemonde.fr --tldr -l fr  # French
-glance https://elpais.com -k -l es      # Spanish
-glance https://example.ht -l ht          # Haitian Creole
+# Read aloud immediately
+glance https://blog.com --read
 ```
 
----
-
-## 🎯 Core Features
-
-### AI Summarization
-
-| Mode | Flag | Description | Best For |
-|------|------|-------------|----------|
-| **Standard** | (default) | Comprehensive 250-400 word summary | Most use cases |
-| **TL;DR** | `-t`, `--tldr` | One powerful sentence (max 25 words) | Quick insights |
-| **Key Points** | `-k`, `--key-points` | 6-10 substantive bullet points | Structured overviews |
-| **ELI5** | `--eli5` | Simple explanation with analogies | Complex topics |
-| **Custom Q&A** | `-q`, `--ask "question"` | Answer specific questions | Targeted information |
-
-**Examples:**
+### Advanced Features
 ```bash
-# Standard summary
-glance https://arxiv.org/abs/2401.12345
-
-# TL;DR for quick scanning
-glance https://techcrunch.com/article --tldr
-
-# Key points for meeting notes
-glance https://docs.example.com --key-points --export notes.md
-
-# ELI5 for learning
-glance https://quantum-computing-paper.com --eli5
+# Screenshot + summary
+glance https://example.com --screenshot page.png --tldr
 
-# Custom questions
-glance https://api-docs.com --ask "How do I authenticate?"
-```
-
-### AI Models
-
-#### 🆓 Local Models (Ollama) – **FREE & Private** (Recommended!)
-```bash
-# Install Ollama first: https://ollama.ai
-ollama pull llama3
-
-# Use with glance (auto-detected as default!)
-glance https://example.com  # Automatically uses Ollama if available
-glance https://example.com --model llama3 --stream
-```
-
-**Popular FREE models:**
-- `llama3` – Meta's flagship (best overall) - **RECOMMENDED**
-- `mistral` – Fast and capable
-- `phi3` – Lightweight, good for resource-constrained
-- `gemma2` – Google's open model
-- `codellama` – Optimized for code
-
-**Cost:** $0.00 forever! Runs on your machine.
-
-#### 💰 Cloud Models (Optional - Costs Money)
+# JavaScript-heavy sites (SPAs)
+glance https://react-app.com --full-render
 
-**OpenAI (Paid):**
-```bash
-export OPENAI_API_KEY=sk-...  # Costs ~$0.0015 per summary
-glance https://example.com --model gpt-4o
-glance https://example.com --model gpt-4o-mini  # Cheaper option
-```
+# Multiple languages
+glance https://lemonde.fr --tldr -l fr  # French
+glance https://elpais.com -k -l es      # Spanish
 
-**Google Gemini (Free tier available, then paid):**
-```bash
-export GEMINI_API_KEY=...
-glance https://example.com --model gemini-2.0-flash-exp  # Has free tier
-glance https://example.com --model gemini-1.5-pro  # Premium
+# Streaming output
+glance https://long-article.com --stream
 ```
 
-**Note:** Glance automatically detects and prefers FREE local models. Use `--prefer-quality` to prioritize paid services.
+---
 
-### Output Formats
+## 🎨 Output Formats
 
 ```bash
-# Terminal (default) - colorful, interactive
+# Terminal (default) - colorful, emoji-rich
 glance https://example.com
 
-# Markdown - for documentation
-glance https://example.com --markdown -o article.md
+# Markdown
+glance https://example.com -o article.md
 
-# JSON - for APIs and scripts
-glance https://example.com --json -o data.json
+# JSON (for scripting)
+glance https://example.com -o data.json
 
-# HTML - embeddable webpage
+# HTML (embeddable)
 glance https://example.com --format html -o page.html
 
-# Plain text - pipeable, script-friendly
+# Plain text (pipeable)
 glance https://example.com --format plain | grep "keyword"
 ```
 
-**Auto-detection:**
-```bash
-# Format auto-detected from extension
-glance url -o output.md    # Markdown
-glance url -o output.json  # JSON
-glance url -o output.html  # HTML
-glance url -o output.txt   # Plain text
-```
+---
 
-### Caching System
+## 💾 Smart Caching
 
-Intelligent caching for speed and cost savings:
+Glance automatically caches results to save time and API costs:
 
 ```bash
-# Cache enabled by default (24h TTL)
+# First request: ~3 seconds
 glance https://example.com
-glance https://example.com  # ⚡ Instant from cache
-
-# View cache statistics
-glance --cache-stats
-# Output:
-#   Total Entries: 234
-#   Total Size: 45.2 MB / 100 MB (45%)
-#   Cache Hits: 892
-#   Cache Misses: 108
-#   Hit Rate: 89.2%
-#   Compression: 72% savings
-
-# Clear expired entries
-glance --cache-cleanup
-
-# Clear all cache
-glance --clear-cache
-
-# Disable cache for specific request
-glance https://example.com --no-cache
-```
-
-**Cache Features:**
-- ✅ Automatic compression (70% space savings)
-- ✅ TTL-based expiration (default 24h)
-- ✅ LRU eviction (max 100MB)
-- ✅ Hit rate tracking
-- ✅ Atomic writes (corruption-proof)
-
-### Content Extraction
 
-Advanced extraction with intelligent content detection:
-
-```bash
-# Extract main content with metadata
-glance https://article.com --metadata
-
-# Extract all links (categorized)
-glance https://example.com --links
-
-# Extract tables from data pages
-glance https://data-table.com --tables
-
-# Get structured metadata (author, dates, etc.)
-glance https://blog.com --metadata --json
-```
-
-**Metadata includes:**
-- Title, description, keywords
-- Author, publication date, publisher
-- Reading time estimation
-- OpenGraph and Twitter Cards
-- Structured data (JSON-LD, Schema.org)
-- Word count, paragraph count
-- Language detection
-
-### Screenshots
-
-```bash
-# Capture screenshot
-glance https://example.com --screenshot page.png
-
-# Custom size and format
-glance https://example.com --screenshot output.jpg --width 1280 --height 720
-
-# Full page screenshot
-glance https://example.com --screenshot full.png --full-page
-```
-
-### Voice/TTS Output
-
-Transform any summary into audio for listening on-the-go:
-
-```bash
-# Read summary aloud (plays immediately)
-glance https://article.com --read
-
-# Save as audio file
-glance https://blog.com --audio-output summary.mp3
-
-# Use specific voice
-glance https://news.com --read --voice alloy
+# Cached request: ~0.1 seconds
+glance https://example.com
 
-# List available voices
-glance --list-voices
+# View cache stats
+glance --cache-stats
 
-# Combine with other features
-glance https://techcrunch.com --tldr --read --emoji
-glance https://paper.com --key-points --audio-output points.mp3
+# Clear cache
+glance --clear-cache
 ```
 
-**Voice Options:**
-- **ElevenLabs** (with API key): Premium natural voices
-  - `alloy` (Rachel) - warm, friendly
-  - `echo` (Clyde) - professional
-  - `nova` (Charlie) - energetic
-  - `shimmer` (Charlotte) - clear, articulate
-  - `onyx` (Callum) - deep, authoritative
-  - `fable` (Gigi) - expressive
-
-- **Local TTS** (no API key needed):
-  - **macOS**: System voices via `say` command
-  - **Windows**: SAPI voices
-  - **Linux**: espeak or festival
-
-**Setup:**
-```bash
-# Optional: For premium ElevenLabs voices
-export ELEVENLABS_API_KEY=your-api-key
-
-# Works without API key using system TTS
-glance https://example.com --read
-```
+**Cache features:**
+- 24-hour TTL (configurable)
+- 70% compression
+- LRU eviction (100MB max)
+- Hit rate tracking
 
 ---
 
-## 🛠️ Advanced Usage
-
-### Streaming Output
+## 🛠️ Common Use Cases
 
-Real-time output for long content:
+### 1. Quick Documentation Lookup
 ```bash
-glance https://long-article.com --stream
+glance https://nextjs.org/docs --ask "How do I use Server Actions?"
 ```
 
-### Custom Timeouts
-
+### 2. Morning News Briefing
 ```bash
-# Increase timeout for slow sites
-glance https://slow-site.com --timeout 60000  # 60 seconds
+glance https://news.ycombinator.com --tldr --read
 ```
 
-### Full Rendering
-
-For JavaScript-heavy sites (SPAs, React apps):
+### 3. Research & Note-Taking
 ```bash
-glance https://react-app.com --full-render
-
-# With custom wait strategy
-glance https://spa.com --full-render --wait-until load
+glance https://research-paper.com --key-points --export notes.md
 ```
 
-### Batch Processing
-
+### 4. Batch Processing
 ```bash
-# Process multiple URLs
 cat urls.txt | while read url; do
-  glance "$url" --tldr --export "summaries/$(echo $url | md5).md"
+  glance "$url" --tldr >> daily-summary.md
 done
 ```
 
-### Piping and Scripting
-
+### 5. API Documentation
 ```bash
-# Extract and grep
-glance https://docs.com --format plain | grep "API"
-
-# Extract links and process
-glance https://example.com --links --json | jq '.[] | .href'
-
-# Chain with other tools
-glance https://article.com | wc -w  # Word count
+glance https://api-docs.com --ask "How do I authenticate?" --format plain
 ```
 
 ---
 
-## 📚 Complete Options
-
-```
-Usage: glance <url> [options]
-
-Core:
-  -t, --tldr                One-line TL;DR (max 25 words)
-  -k, --key-points          6-10 key insights as bullet points
-      --eli5                Explain like I'm 5 (simple with analogies)
-
-Questions:
-  -q, --ask "question"      Ask anything about the page
-
-AI Models:
-  -m, --model <name>        Model to use (default: gpt-4o-mini)
-                            OpenAI: gpt-4o, gpt-4o-mini
-                            Gemini: gemini-2.0-flash-exp, gemini-1.5-pro
-                            Ollama: llama3, mistral, phi3, gemma2, etc.
-  -l, --language <code>     Output language: en, fr, es, ht (default: en)
-      --max-tokens <n>      Limit output length
-  -e, --emoji               Add relevant emojis
-
-Voice/TTS:
-      --read, --speak       Read summary aloud (text-to-speech)
-      --voice <name/id>     Voice to use (alloy, echo, nova, etc.)
-      --audio-output <file> Save audio to MP3 file
-      --list-voices         List available voices
-
-Output:
-      --markdown            Markdown format
-  -j, --json                JSON format
-      --format <type>       Output format: terminal, markdown, json, html, plain
-  -o, --export <file>       Save to file (auto-detects format from extension)
-      --stream              Live streaming output
-      --include-metadata    Include full metadata in output
-
-Cache:
-  -c, --cache               Enable cache (default: true)
-      --no-cache            Disable cache for this request
-      --clear-cache         Clear all cache
-      --cache-cleanup       Remove expired entries
-      --cache-stats         Show cache statistics
-
-Advanced:
-  -r, --full-render         Render JavaScript (slower, for SPAs)
-      --timeout <ms>        Request timeout in milliseconds
-      --screenshot <file>   Save screenshot
-      --width <n>           Screenshot width (default: 1920)
-      --height <n>          Screenshot height (default: 1080)
-      --raw-html            Print raw HTML
-      --links               List all links
-      --tables              Extract tables
-      --metadata            Show page metadata
-      --list-models         List available Ollama models
-
-Debug:
-  -v, --verbose             Verbose logging (shows performance breakdown)
-  -d, --dry-run             Preview content without AI processing
-
-Help:
-  -h, --help                Show this help
-  -V, --version             Show version number
+## 📚 All Options
+
+```bash
+# Summary modes
+glance <url>                  # Standard summary
+glance <url> --tldr           # One sentence
+glance <url> --key-points     # Bullet points
+glance <url> --eli5           # Simple explanation
+glance <url> --ask "question" # Custom Q&A
+
+# AI models
+--model <name>                # llama3, gpt-4o-mini, gemini-2.0-flash-exp
+--list-models                 # Show available Ollama models
+--check-services              # Show available AI/voice services
+--free-only                   # Never use paid APIs
+--prefer-quality              # Use premium services when available
+
+# Output
+-o, --export <file>           # Save to file (auto-detects format)
+--markdown                    # Markdown format
+--json                        # JSON format
+--format <type>               # terminal, markdown, json, html, plain
+--stream                      # Live streaming output
+
+# Voice/Audio
+--read, --speak               # Read aloud (text-to-speech)
+--audio-output <file>         # Save as MP3
+--voice <name>                # Choose voice (alloy, echo, nova, etc.)
+--list-voices                 # Show available voices
+
+# Advanced
+--screenshot <file>           # Capture screenshot
+--full-render                 # Render JavaScript (for SPAs)
+--metadata                    # Show page metadata
+--links                       # Extract all links
+--tables                      # Extract tables
+-l, --language <code>         # Output language (en, fr, es, ht)
+-e, --emoji                   # Add emojis
+
+# Cache
+--cache-stats                 # Show cache statistics
+--cache-cleanup               # Remove expired entries
+--clear-cache                 # Clear all cache
+--no-cache                    # Skip cache for this request
+
+# Other
+-v, --verbose                 # Show detailed logs
+-h, --help                    # Show help
+-V, --version                 # Show version
 ```
 
 ---
@@ -526,349 +274,118 @@ Help:
 ### Environment Variables
 
 ```bash
-# AI Providers
+# AI Providers (optional)
 export OPENAI_API_KEY=sk-...
 export GEMINI_API_KEY=...
-export GOOGLE_API_KEY=...  # Alternative for Gemini
 
-# Voice/TTS (optional)
-export ELEVENLABS_API_KEY=...  # For premium voices
+# Voice (optional)
+export ELEVENLABS_API_KEY=...
 
-# Ollama (optional, default shown)
+# Ollama (optional, auto-detected)
 export OLLAMA_ENDPOINT=http://localhost:11434
 
 # Cache (optional)
 export GLANCE_CACHE_DIR=~/.glance/cache
 export GLANCE_CACHE_TTL=86400000  # 24h in ms
-export GLANCE_MAX_CACHE_SIZE=104857600  # 100MB
-```
-
-### Configuration File (coming soon)
-
-`.glancerc` support planned for v1.1:
-```json
-{
-  "defaultModel": "gpt-4o-mini",
-  "defaultLanguage": "en",
-  "cacheEnabled": true,
-  "cacheTTL": 86400000,
-  "maxCacheSize": 104857600
-}
 ```
 
 ---
 
-## 📖 Examples
-
-### Content Summarization
-
-```bash
-# Summarize a blog post
-glance https://blog.example.com/post
-
-# Get key points from documentation
-glance https://docs.python.org/3/tutorial/ --key-points
-
-# Quick TL;DR of news article
-glance https://techcrunch.com/article --tldr --emoji
-```
+## 🚀 Examples
 
 ### Learning & Research
-
 ```bash
-# Understand a complex topic
-glance https://quantum-computing.com --eli5
+# Understand a new framework
+glance https://nextjs.org/docs --eli5
 
-# Answer specific questions
-glance https://react.dev --ask "What are hooks and when should I use them?"
+# Quick reference
+glance https://nextjs.org/docs --key-points --export nextjs-notes.md
 
-# Extract structured data
-glance https://research-paper.com --metadata --json -o paper.json
+# Specific questions
+glance https://nextjs.org/docs --ask "What's the difference between pages and app router?"
 ```
 
-### Documentation & Notes
-
+### Daily Workflow
 ```bash
-# Export documentation to Markdown
-glance https://api-docs.com --markdown -o api-reference.md
+# Morning news
+glance https://news.ycombinator.com --tldr --read
 
-# Create meeting notes from recorded session
-glance https://meeting-recording.com --key-points --export notes.md
+# Documentation lookup
+glance https://nextjs.org/docs/app/api-reference/functions/fetch --ask "How do I revalidate?"
 
-# Extract code examples
-glance https://tutorial.com --format plain | grep -A 10 "```"
+# Code examples
+glance https://tutorial.com --format plain | grep -A 10 "example"
 ```
 
-### Monitoring & Automation
-
+### Content Creation
 ```bash
-# Check website changes (with caching)
-glance https://status.example.com --tldr
-
-# Automated reporting
-#!/bin/bash
-for url in $(cat monitoring.txt); do
-  glance "$url" --tldr --format plain >> daily-report.txt
-done
+# Extract for blog post
+glance https://source-article.com --key-points --export research.md
 
-# API integration
-curl -s "https://api.example.com/articles" | jq -r '.[]|.url' | \
-  xargs -I {} glance {} --key-points --json
+# Multilingual content
+glance https://lemonde.fr --tldr -l fr --export french-news.md
 ```
 
 ---
 
-## 🎓 Best Practices
+## 🎓 Pro Tips
 
-### Choosing the Right Model
-
-**For speed and cost:**
+**1. Use local AI for daily tasks:**
 ```bash
-glance url --model gpt-4o-mini  # Fast, cheap, good quality
-glance url --model llama3       # Free, local, private
+glance https://example.com --model llama3  # Free, fast, private
 ```
 
-**For highest quality:**
+**2. Cache speeds up repeated requests:**
 ```bash
-glance url --model gpt-4o       # Best reasoning
-glance url --model gemini-1.5-pro  # Long context
+glance https://docs.com  # First: 3s
+glance https://docs.com  # Cached: 0.1s
 ```
 
-**For code/technical content:**
+**3. Pipe output for scripting:**
 ```bash
-glance url --model codellama    # Optimized for code
+glance https://example.com --format plain | grep "keyword" | wc -l
 ```
 
-### Optimizing Performance
-
+**4. Save audio for commute:**
 ```bash
-# Use cache for repeated requests
-glance url  # First time: 3s
-glance url  # Cached: 0.1s
-
-# Use streaming for long content
-glance long-url --stream
-
-# Skip JavaScript rendering if not needed
-glance url  # Fast (static fetch)
-glance spa-url --full-render  # Slower but complete
+glance https://article.com --tldr --audio-output commute.mp3
 ```
 
-### Error Handling
-
+**5. Use streaming for long content:**
 ```bash
-# Use verbose mode for debugging
-glance url --verbose
-
-# Check URL before processing
-glance url --dry-run
+glance https://long-article.com --stream
 ```
 
 ---
 
 ## 🤝 Contributing
 
-We welcome contributions! Here's how to get started:
-
-### Development Setup
+Contributions welcome! Check out our [Contributing Guide](CONTRIBUTING.md).
 
 ```bash
-# Clone repository
 git clone https://github.com/jkenley/glance-cli.git
 cd glance-cli
-
-# Install dependencies
 bun install
-
-# Run locally
 bun dev https://example.com
-
-# Run tests
-bun test
-
-# Build
-bun build
 ```
 
-### Contribution Guidelines
-
-1. **Open an issue** for major changes
-2. **Follow code style** (TypeScript, ESLint)
-3. **Add tests** for new features
-4. **Update documentation** as needed
-5. **Write clear commit messages**
-
-See [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines.
-
----
-
-## 🏗️ Architecture
-
-### Core Components
-
-- **`cli.ts`** - Command-line interface with validation and error handling
-- **`fetcher.ts`** - Page fetching with retry logic and timeout protection
-- **`extractor.ts`** - Intelligent content extraction with scoring algorithm
-- **`summarizer.ts`** - Expert-engineered AI prompts for superior results
-- **`formatter.ts`** - Multi-format output (terminal, markdown, JSON, HTML)
-- **`cache.ts`** - Smart caching with TTL, compression, and LRU eviction
-- **`screenshot.ts`** - Screenshot capture with error recovery
-
-### Production Features
-
-- ✅ **Comprehensive error handling** with categorized errors and hints
-- ✅ **Automatic retries** with exponential backoff for transient failures
-- ✅ **Timeout protection** to prevent infinite hangs
-- ✅ **Resource cleanup** to prevent memory leaks
-- ✅ **Input validation** to catch issues early
-- ✅ **Graceful degradation** on partial failures
-- ✅ **Performance tracking** with detailed breakdowns
-
----
-
-## 📊 Performance
-
-### Benchmarks
-
-| Operation | Time | Notes |
-|-----------|------|-------|
-| Simple fetch | 0.5-1s | Static page |
-| Full render | 2-4s | JavaScript-heavy SPA |
-| AI summary (cloud) | 2-5s | Depends on model |
-| AI summary (local) | 3-8s | Depends on hardware |
-| Cache hit | <0.1s | Nearly instant |
-
-### Optimization Tips
-
-1. **Use cache** - Reduces API costs and improves speed
-2. **Choose appropriate model** - `gpt-4o-mini` is 10x faster than `gpt-4o`
-3. **Skip full-render** if not needed - 4x faster for static sites
-4. **Use streaming** for long content - Better perceived performance
-5. **Batch processing** - Process multiple URLs efficiently
-
----
-
-## 🔒 Security
-
-### Reporting Vulnerabilities
-
-See [SECURITY.md](SECURITY.md) for our security policy.
-
-**Report security issues to:** alo@jkenley.me with subject `[Security] glance-cli`
-
-### Security Features
-
-- ✅ **Input sanitization** - All inputs validated and sanitized
-- ✅ **No shell injection** - Safe programmatic operations
-- ✅ **Content-type validation** - Only processes valid HTML
-- ✅ **Size limits** - Prevents memory exhaustion attacks
-- ✅ **Timeout protection** - Prevents denial of service
-- ✅ **Secure hashing** - SHA-256 instead of MD5
-
----
-
-## 📝 Changelog
-
-See [CHANGELOG.md](CHANGELOG.md) for version history and updates.
-
-### Latest Release (v0.7.0)
-
-**Major Improvements:**
-- 🎯 Expert-level prompt engineering (10x better AI results)
-- 🛡️ Production-grade error handling and reliability
-- 💾 Intelligent caching with TTL, compression, and LRU eviction
-- 📊 Rich content extraction with 20+ metadata fields
-- 🎨 Multiple output formats (HTML, plain text)
-- ⚡ Performance optimizations and memory leak fixes
-
-See full changelog for all updates.
-
----
-
-## 🌟 Roadmap
-
-### v1.0 (Stable Release)
-- [ ] Comprehensive test suite (>80% coverage)
-- [ ] Configuration file support (`.glancerc`)
-- [ ] Plugin system for extensibility
-- [ ] Web UI for easier interaction
-- [ ] Additional output formats (PDF, EPUB)
-
-### v1.1 (Enhanced Features)
-- [ ] Batch processing with progress bars
-- [ ] Interactive mode for exploration
-- [ ] Custom prompt templates
-- [ ] Integration with note-taking apps (Notion, Obsidian)
-- [ ] Browser extension
-
-### Future
-- [ ] Cloud sync for cache
-- [ ] Analytics dashboard
-- [ ] Collaborative features
-- [ ] Mobile app
-
-Vote on features: [GitHub Discussions](https://github.com/jkenley/glance-cli/discussions)
-
----
-
-## ❓ FAQ
-
-**Q: Which AI model should I use?**
-A: For most cases, `gpt-4o-mini` (default) offers the best balance of speed, cost, and quality. For privacy or offline use, try `llama3` with Ollama.
-
-**Q: How does caching work?**
-A: Responses are cached for 24 hours by default. The cache uses intelligent compression and LRU eviction to stay under 100MB.
-
-**Q: Can I use glance without an API key?**
-A: Yes! Install Ollama and use local models like `llama3` or `mistral` for completely free, private operation.
-
-**Q: Does glance work with paywalled content?**
-A: No, glance cannot access content behind authentication or paywalls.
-
-**Q: How accurate are the summaries?**
-A: glance uses expert-engineered prompts with anti-hallucination measures. Summaries are based strictly on page content, but AI can still make mistakes. Always verify critical information.
-
-**Q: Can I customize the prompts?**
-A: Custom prompt templates are planned for v1.1. Currently, you can use different modes (TL;DR, key points, ELI5) or ask custom questions.
-
-**Q: Is my data sent to third parties?**
-A: When using cloud models (OpenAI, Gemini), page content is sent to those providers for processing. With local models (Ollama), everything stays on your machine.
-
----
-
-## 🙏 Acknowledgments
-
-Built with these amazing open-source projects:
-- [Bun](https://bun.sh) - Lightning-fast JavaScript runtime
-- [Cheerio](https://cheerio.js.org) - Fast, flexible HTML parsing
-- [Puppeteer](https://pptr.dev) - Headless Chrome automation
-- [OpenAI](https://openai.com) - GPT models
-- [Google Gemini](https://deepmind.google/technologies/gemini/) - Gemini models
-- [Ollama](https://ollama.ai) - Local LLM infrastructure
-
-Special thanks to all [contributors](https://github.com/jkenley/glance-cli/graphs/contributors)!
-
 ---
 
 ## 📄 License
 
 MIT License - see [LICENSE](LICENSE) for details.
 
-Copyright © 2026 Kenley Jean-Louis
-
 ---
 
-## 📞 Contact & Support
+## 🔗 Links
 
-- **Author**: Kenley Jean-Louis
-- **Email**: alo@jkenley.me
-- **GitHub**: https://github.com/jkenley
-- **Project**: https://github.com/jkenley/glance-cli
-- **Issues**: https://github.com/jkenley/glance-cli/issues
-- **Discussions**: https://github.com/jkenley/glance-cli/discussions
+- **GitHub**: [github.com/jkenley/glance-cli](https://github.com/jkenley/glance-cli)
+- **NPM**: [npmjs.com/package/glance-cli](https://www.npmjs.com/package/glance-cli)
+- **Issues**: [github.com/jkenley/glance-cli/issues](https://github.com/jkenley/glance-cli/issues)
+- **Changelog**: [CHANGELOG.md](CHANGELOG.md)
 
 ---
 
-**Built with ❤️ for developers, researchers, and curious minds.**
+**Built with ❤️ by [Kenley Jean](https://github.com/jkenley)**
 
 *Star ⭐ the repo if you find it useful!*

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "glance-cli",
-    "version": "0.8.0",
+    "version": "0.8.2",
     "description": "AI-powered web reader for your terminal – Fetch, extract, and summarize webpages with local or cloud AI",
     "keywords": [
         "cli",