There is a silent monopoly in the AI code editor space, and it relies on a concept known as the Token Tax.
Right now, the average senior developer already pays $20/month for a baseline consumer subscription like Claude Pro or ChatGPT Plus. But when they download a proprietary "AI-first" IDE like Cursor or Windsurf, they are immediately coerced into paying another$20/month for a "Pro" IDE tier.
Worse yet, this IDE subscription doesn't give you unlimited access. It gives you an arbitrary pool of "fast requests" or "credits" (usually maxing out at around 500 prompts).
Agentic workflows require passing massive context windows (100k+ tokens) back and forth. Power users regularly burn through their entire monthly pool of 500 IDE credits within the first 48 hours of heavy feature development.
Once you exhaust those credits, standard IDEs throttle your compute speed to a crawl unless you purchase high-margin "top-up" API tokens directly from them, effectively forcing you into a retail pricing model for wholesale computer logic.
Bring Your Own Subscription (BYOS) removes the editor as a token reseller. You connect supported provider credentials or subscriptions directly, while usage limits and pricing continue to come from the provider and plan you chose.
Glass was engineered so it does not resell your AI usage at a markup.
Glass runs on a Bring Your Own Subscription model. You connect supported Anthropic, Google, or OpenAI access directly in the application.
That keeps provider pricing and usage limits between you and the provider, instead of adding a second token pool or editor markup.