Make it Phanthy.

Download the PhanthyCode or PhanthyWork desktop app for your system, or install the PhanthyCode CLI with the official install script. Log in with the same Phanthy account, then start working from the desktop or terminal.

Phanthy-Fast, Phanthy-Pro, and Phanthy-Ultra do not map permanently to any one AI model. They are three tiers of intelligent model service offered by PhanthyCode. Based on the real-world capabilities, pricing, and availability of leading AI models worldwide, PhanthyCode uses intelligent model routing to dynamically select a suitable model: Phanthy-Fast prioritizes cost-effective models with fast response times; Phanthy-Pro prioritizes stronger frontier models with a better balance of performance and cost; Phanthy-Ultra prioritizes the most capable models currently available for highly demanding tasks. As model capabilities evolve, the actual models behind these three tiers may also change. You do not need to keep track of which model is newest or best suited to the current task. Just choose the right tier and leave model selection to PhanthyCode.

PhanthyCode uses points as a unified measure of model usage allowance. 1,000 points = US$1 of model allowance. For example, 10,000 points in your account are equivalent to US$10 of model usage allowance. Actual point consumption is calculated from the price of the model used and the actual token usage. Using points as a unified unit makes usage and cost across different models easier to understand and lets you switch freely between models.