FAQ
Short answers about licensing, data, model quality, Studio, platforms and how BdebTech AI IDE compares to cloud tools.
Is it free?
Yes. The IDE, the extension, the daemon and the two free models have no cost and no usage limits. Studio is an optional paid plan for larger models and support.
Does my code go anywhere?
No. The model runs on your machine; the editor talks to it over 127.0.0.1. There is no telemetry. See Offline mode and privacy.
Can I use it for commercial work?
Yes. The BdebTech components are Apache-2.0, VSCodium is MIT, llama.cpp is MIT, and Qwen2.5-Coder weights are Apache-2.0. Output you generate is yours.
How good is the AI?
The Starter model is excellent at completing routine code and answering focused questions, and weak at open-ended design. Lite+ is a clear step up. Studio models (7B-32B) are competitive with cloud assistants for most day-to-day coding when run on a GPU. See Models and hardware.
Why is the default model so small?
So that the very first run works instantly on any laptop. You are shown bigger options as soon as your hardware is profiled.
Can I use my own GGUF model?
Yes. Model switcher → Add a custom GGUF…, or bdebd models add <path or URL>. See Models and hardware.
Does it index my whole repository?
Yes, if you let it. The IDE offers to build a local semantic index of each folder you open; chat then has an @codebase toggle and autocomplete receives related snippets from other files. See Repository awareness. Everything stays on your machine.
What does Studio cost and how do I pay?
$200 per month, prepaid, with quarterly and yearly discounts. UPI in India; USDT, USDC or Binance Pay worldwide. See Pricing.
Which platforms?
Windows 10/11 x64 today. Linux and macOS builds are next; the engine downloads for them already exist in the registry.
Is the editor a fork of VS Code?
It is stock VSCodium (the MIT-licensed community build of Code - OSS) with our extension pre-installed and light branding. We do not maintain a fork, which means you keep getting upstream editor fixes.