JetBrains announced on October 16 that its local AI coding tool Junie Local has day-one support for Nvidia's RTX Spark platform, and that Windows support has moved from experimental to stable. On eligible machines the tool can read code, edit files and run tests; JetBrains says it needs no sign-up or subscription, charges no token fees, imposes no usage quotas and keeps code off the cloud. In initial testing JetBrains measured generation speeds of up to 55 tokens per second on RTX Spark. Requirements include at least 24GB of VRAM on an Ampere-or-newer RTX card (RTX 3090, 4090 and the 32GB 5090 are listed), Windows 11, driver version 581 or later and about 20GB of disk space; the installer checks the GPU and driver, deploys the model and scales context to available VRAM, with a choice of Qwen3.6-27B or a JetBrains-built merged model.