0.1.5 is the first ORKA VOICE preview that’s signed and notarized. That’s the unglamorous milestone that turns a project you built for yourself into something other people can actually install without fighting Gatekeeper. It ships with a branded, drag-to-Applications installer, so setup is the thing every Mac user already knows how to do.
The rest of this release is about the moment between pressing the key and seeing your words. We made Fn activation faster, and taught it to handle Bluetooth audio — including AirPods — instead of stumbling when the input device isn’t the built-in mic.
Processing now stays visible until the transcript is actually pasted. No more wondering whether it heard you: the status pill holds through listening, processing, and completion, so the app’s state always matches what’s about to land at your cursor. When something does go wrong — a crash, a stall, a microphone that won’t start — there are now privacy-conscious diagnostics to make it debuggable, without shipping your audio or transcripts anywhere.
Still true under the hood
None of this changes what ORKA VOICE is. Transcripts still get an on-device AI cleanup that strips filler and fixes punctuation — and quietly falls back to the raw text if the cleanup would mangle it. Your clipboard is snapshotted before a paste and restored right after, so your own copy-paste flow is left untouched. A custom dictionary keeps names and jargon spelled the way you mean them, and up to 500 entries of local, searchable history live as JSON on your Mac — nothing leaves the machine.
App updates come through a signed Sparkle feed, kept deliberately separate from the website, so a marketing change can never masquerade as an app update. That’s the shape of it: a preview, but a real one.