Media
1 item
Description
Real-time token counts, costs & energy usage on ChatGPT, Claude, Gemini, Grok, DeepSeek + 25 more. Free, open source. Know exactly what every AI message costs — before you hit send. Tokenizer adds a small, non-intrusive overlay to your favorite AI chat interfaces. As you type, it shows a live token count, dollar cost estimate (input and output), and a running session total. No setup required — just install and start chatting. KEY FEATURES Live token counter that updates with every keystroke, so you always know the size of your prompt before sending it. Cost estimates based on current model pricing, broken down by input tokens, output tokens, and cumulative session spend. Cross-model comparison lets you see what the same prompt would cost across different AI models side by side. Word-level token breakdown with color-coded chips showing which words use the most tokens — helpful for optimizing long prompts. Energy usage estimates showing approximate watt-hours and CO2 per prompt, for users who care about the environmental impact of AI. Works with local AI tools running on localhost, including popular self-hosted setups. PRIVACY Everything runs 100% locally in your browser. There is no data collection, no accounts, no external servers. All token counting happens entirely on your device. Open source under the MIT License at github.com/gyeningcorp/tokenizer