We built DictaFlow because we were frustrated with existing dictation tools that always run in the background and don't work well in code editors or remote desktops. DictaFlow is a lightweight, privacy-first speech-to-text tool for Windows that only listens when you hold down the activation key and inserts text when you release. It works in any application (including terminals, IDEs, VDI/Citrix sessions), does context-aware formatting for code, math, and natural language, and processes nothing when idle. We don't store audio or transcripts on our servers. There's a free tier as well as a paid plan.

We'd love feedback from the HN community on how it performs, any features you'd like, or issues you encounter.