UltraToken v1.0.3

UltraToken is a CLI utility that replicates TikTokens BPE tokenizer, and utilizes OpenAIs o200K Harmony encodings for fast, precise token cost estimation.

Features

• Zero external dependencies - entirely self contained
• Complete BPE Implementation - Full byte-pair encoding algorithm
• Embedded o200k Vocabulary - 200k token vocabulary included
• Accurate Regex Splitting - Matches OpenAI's text segmentation
• Optimized Token Mapping - Fast lookups using Map structures
• Special Token Support - Handles control tokens correctly
• Batch Processing

Installation

Install UltraToken with npm

  npm i ultratoken

Or clone and run locally:

git clone https://github.com/TrintechResearch/UltraToken.git
cd UltraToken
npm install -g .

Usage

Command Overview

Command	Description
ultratoken	Start interactive mode
ultratoken <text>	Get token count for text
ultratoken economy <file.md>	Process word list file
ultratoken jump	Exit the program
ultratoken --help	Show help information
ultratoken --version	Show version information

Interactive Mode

Start the interactive session:

ultratoken

🚀 UltraToken TikToken Utility
Interactive Mode - Type words to get token counts
Commands: "jump" to exit, "help" for help

ultratoken hello world
"hello world" = 2 tokens

ultratoken programming
"programming" = 2 tokens

ultratoken The quick brown fox
"The quick brown fox" = 4 tokens

ultratoken jump
UltraToken terminated. Goodbye!

Single Text Analysis

Get token count for any text:

ultratoken "machine learning"
# Output:
# Word: machine learning
# Tokens: 2

ultratoken "The quick brown fox jumps over the lazy dog"
# Output:
# Word: The quick brown fox jumps over the lazy dog
# Tokens: 9

Economy Mode - Batch Processing

Process a file containing a list of words. UltraToken will append the token count to each line:

Input file (words.txt):

hello world
artificial intelligence
machine learning
programming
natural language processing

Command:

ultratoken economy words.txt

Output file (words.txt) after processing:

hello world 2
artificial intelligence 3
machine learning 2
programming 2
natural language processing 4

Documentation

For detailed documentation and advanced features, see Documentation.md

License

MIT

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgements

• Civil McKnight - Project Ultra
• TrinityAI
• Project Ultra

About

UltraToken is developed by TrinityAI Research, specializing in advanced LLM development.

🔗 Links