I built it as a personal project at first, but want the site to be as transparent as possible. So I made one that shows you exactly which model is responding, lets you watch its reasoning stream live and you can see the tokens counter right in the footer real time. Text only output as of now.

What makes it different: 📍35+ Frontier Models: No tab-switching. Toggle between the latest GPTs, Claude Sonnet 4.5, Llama 4 Scout, DeepSeek, Grok, Gemini, Mistral and more, all in one click. 📍Token Pool Separation: Tokens are separated into Casual and Pro Token pools. Casual models use Casual tokens, Pro and Web Search models use Pro tokens. This allows you to optimize your token usage based on model type. The Tokens are further Divided into Input and Output for each pool. 📍8M Monthly Tokens for $15/mo: 8M tokens per month, split into : 5M Casual Input / 1.5M Casual Output, 1M Pro Input / 500k Pro Output, 100 Web Searches (tokens will be deducted from pro pool) 📍Model Weight: Each model will have a model weight. A model with 0.5x weight will consume only half as many tokens where as a model with 2x weight will consume tokens at twice the rate. 📍Token Transfer System: You can transfer tokens between Input and Output within same pool after a conversion rate is applied, i.e., between Casual Input and Output, and between Pro Input and Output. 📍Real-time Reasoning: Watch the model's full thought process unfold alongside the answer (when supported). 📍Midchat Model Change: You can switch models in the middle of a chat at any given time. 📍Context Window: We have context windows ranging from 32k till 128k tokens depending on the model in use. 🔒Data Encryption: We encrypt Prompt Text, AI Response and AI reasoning using AES 256 CBC before they are saved in database. 🗑️Data Deletion: Hard delete policy is followed once you click on delete chat option. The Solo Dev Promise: No marketing team, no fancy office. Just me, my laptop, and commitment to building something useful.