refactor: improve demo with better branding and features

- Merged best features from demo.vhs into scripts/demo.tape
- Added title screen with project branding
- Multiple output formats (GIF, MP4, WebM)
- Increased dimensions to 1200x800 for better readability
- Changed to Dracula theme for better aesthetics
- Added verbose encoding demonstration
- Added special tokens decode with --show-special
- Added JSON output with jq for token counting
- Added help command demonstrations
- Added nice ending with thanks and GitHub link
- Removed old demo.vhs to avoid confusion

The demo now provides a comprehensive showcase of tokenizer's features
with better visual presentation and flow.

Author: jackspirou

docs/demo.gif

@@ -1,46 +1,88 @@
-# VHS tape file for generating tokenizer demo GIF
+# VHS tape file for generating tokenizer demo
 # Run with: vhs scripts/demo.tape
 
 # Setup
 Output docs/demo.gif
+Output docs/demo.mp4
+Output docs/demo.webm
+
 Set FontSize 14
-Set Width 800
-Set Height 600
-Set Theme "Monokai Pro"
-
-# Show installation command (but don't run it)
-Type "# Install Tokenizer CLI" Sleep 500ms Enter
-Sleep 1s
-Type "# brew install agentstation/tap/tokenizer" Sleep 500ms Enter
+Set Width 1200
+Set Height 800
+Set Theme "Dracula"
+
+# Title screen
+Type "# Tokenizer - High-performance tokenizer implementations in Go" Sleep 500ms Enter
 Sleep 2s
+Clear
+
+# Install tokenizer
+Type "# Install via Homebrew" Sleep 500ms Enter
+Sleep 500ms
+Type "brew install agentstation/tap/tokenizer" Sleep 500ms Enter
+Sleep 5s
 
 Type "# Check version" Sleep 500ms Enter
 Type "tokenizer version" Sleep 500ms Enter
 Sleep 2s
 
-Ctrl+L
+Clear
 
-Type "# Encode text with Llama 3" Sleep 500ms Enter
-Sleep 1s
+# Basic encoding/decoding
+Type "# Encode text to tokens" Sleep 500ms Enter
+Sleep 500ms
 Type "tokenizer llama3 'Hello, world!'" Sleep 500ms Enter
 Sleep 2s
 
 Type "# Decode tokens back to text" Sleep 500ms Enter
-Type "tokenizer llama3 decode 128000 9906 11 1917 0 128001" Sleep 500ms Enter
+Type "tokenizer llama3 decode 128000 9906 11 1917 59 0 128001" Sleep 500ms Enter
+Sleep 2s
+
+Type "# Decode with special tokens shown" Sleep 500ms Enter
+Type "tokenizer llama3 decode --show-special 128000 9906 11 1917 59 0 128001" Sleep 500ms Enter
 Sleep 2s
 
-Ctrl+L
+Clear
 
-Type "# Get tokenizer info" Sleep 500ms Enter
+# Advanced features
+Type "# Encode with verbose output" Sleep 500ms Enter
+Type "tokenizer llama3 encode --verbose 'The quick brown fox jumps over the lazy dog.'" Sleep 500ms Enter
+Sleep 3s
+
+Clear
+
+# Tokenizer info
+Type "# Get tokenizer information" Sleep 500ms Enter
 Type "tokenizer llama3 info | head -15" Sleep 500ms Enter
 Sleep 3s
 
-Ctrl+L
+Clear
 
-Type "# Stream processing" Sleep 500ms Enter
-Type "echo 'The quick brown fox' | tokenizer llama3 stream" Sleep 500ms Enter
+# Pipe and streaming
+Type "# Stream text from stdin" Sleep 500ms Enter
+Type "echo 'Tokenizers are amazing!' | tokenizer llama3" Sleep 500ms Enter
 Sleep 2s
 
-Type "# Direct encoding (shorthand)" Sleep 500ms Enter
-Type "tokenizer llama3 'I love tokenizers!'" Sleep 500ms Enter
-Sleep 2s
+Type "# Count tokens using JSON output" Sleep 500ms Enter
+Type "echo 'This is a test of the tokenizer.' | tokenizer llama3 encode --format json | jq '.token_count'" Sleep 500ms Enter
+Sleep 2s
+
+Clear
+
+# Help commands
+Type "# Get help and see all options" Sleep 500ms Enter
+Type "tokenizer --help" Sleep 500ms Enter
+Sleep 3s
+
+Clear
+
+Type "# See llama3 specific options" Sleep 500ms Enter
+Type "tokenizer llama3 --help" Sleep 500ms Enter
+Sleep 3s
+
+Clear
+
+# Final message
+Type "# Thanks for watching! 🚀" Sleep 500ms Enter
+Type "# Learn more at: https://github.com/agentstation/tokenizer" Sleep 500ms Enter
+Sleep 3s