VTX Macro’s latest update focuses on helping users better understand what the AI is doing. On History and Analysis pages, users can now choose to display captured reasoning traces when they’re available. That gives more visibility into both primary and review-stage outputs, with cleaner formatting that makes detailed traces easier to read. The assistant and analysis views also preserve original model responses more consistently, so decision context appears more clearly across the app.

This release also strengthens local AI reliability. Device-stored local model settings now preserve server-type details more accurately, which helps prevent setup mismatches for local providers such as Ollama-style connections and OpenAI-compatible endpoints. WebGPU model availability is also more resilient, making model downloads and selection less likely to get blocked by stale error state when the catalog itself is healthy. The overall result is a local AI setup that feels steadier and more predictable.

For users running client-side trading workflows, the runtime now handles changing session and contract conditions more gracefully. Instead of failing hard when a cycle becomes stale, the system discards the outdated cycle and continues forward with refreshed runtime state. That should reduce unnecessary interruptions and make longer-running local sessions feel smoother and more dependable.