--- title: "VoiceMe 1.0: Live Subtitles for Anything" description: "VoiceMe 1.0 turns the sound your computer plays into live subtitles, typed text or a transcript — recognised on your own machine. Captions in about a second." url: "https://royalsoftworks.com/blog/voiceme-1-0-live-subtitles-for-anything/" source: "https://royalsoftworks.com" format: "markdown" note: "Markdown rendering of the HTML page at `url`. Same content, same canonical URL." --- [All posts](https://royalsoftworks.com/blog/) VoiceMe # VoiceMe 1.0: Live Subtitles for Whatever Your Computer Is Playing October 5, 2026 2 min read Royal SoftWorks voiceme changelog release-notes **TL;DR —**VoiceMe 1.0 is the first public release, for Windows 10 and 11 as an installer or a portable .exe. It listens to what your speakers play — not the microphone — and turns it into live subtitles on screen, text typed into another app, or a saved transcript, using whisper.cpp on your own machine. Captions appear in about a second. Everything is free except that the subtitle overlay runs an hour at a time; Pro removes that for €20 once. Plenty of tools turn your _own voice_ into text. VoiceMe does the other thing: it listens to whatever your computer is **playing** — a meeting, a lecture video, a podcast, a stream — and turns it into words. 1.0 is its first public release, on the [download page](https://royalsoftworks.com/download/#voiceme) for Windows. ## Three things it does with the words - **Live subtitles.** A caption window floats over everything else. For streaming it can sit on a chroma-key background so OBS or vMix can key it out; it can be click-through, and it can be kept out of screen captures when the subtitles are for you, not the audience. - **Typing into another app.** VoiceMe can type what it hears into the window you are working in — notes, a document, a chat. Because that touches someone else's application, it is off by default, needs an explicit arm, has a list of apps it must never type into; with that list set, it refuses a window it cannot identify. - **A transcript.** Every finished sentence is saved, restored next launch, and exportable as plain text, Markdown or JSON Lines. ## On your own machine Recognition is whisper.cpp, running locally. The recognition engine and the model are both downloaded from inside the app, each only when you click, and the model list is ranked against a scan of your hardware so the recommendation fits the computer it is going to run on. Nothing is uploaded: no audio, no text, no telemetry. ## Captions in about a second Speech recognition usually waits for the end of a sentence before showing anything, which on continuous speech means waiting a long time. In 1.0 the sentence being spoken is transcribed as it happens, about twice a second. The part that has settled is shown solid and the part still moving is dimmed, and when the sentence ends its finished transcription replaces the preview. Live text is display-only: it is never typed, saved or exported. We also sized the engine's work to the audio actually present instead of the thirty seconds it normally pads every clip to, which roughly halves the time a live caption takes to appear. ## Fixed before 1.0 A correctness audit before release found and fixed a handful of things worth knowing were fixed: overlapping sentences could corrupt the clipboard while typing, restoring the clipboard could wipe an image you had copied, the block list failed open on an unidentifiable window, and browser noise suppression was muting the very audio being transcribed. The list is on the [changelog](https://royalsoftworks.com/products/voiceme/changelog/). ## What it costs Listening, transcription, typing and transcripts are free and unlimited. The one limit is on the subtitle overlay, which runs an hour at a time with a five-minute pause in between. **Pro is €20 once, VAT included** — one computer, no subscription, future updates included. See [pricing](https://royalsoftworks.com/pricing/) or the [VoiceMe page](https://royalsoftworks.com/products/voiceme/). ## Frequently asked questions Does VoiceMe use my microphone? No. It captures the audio your computer is playing — the same sound you hear from the speakers or headphones — so it can subtitle a call, a video or a stream without anyone speaking into a microphone. Is anything uploaded? No audio and no text. Recognition runs on your own machine with whisper.cpp. The engine and the model are downloaded from inside the app when you click, the licence is checked when you activate, and the app looks for updates — nothing else goes out. How fast are the captions? About a second. The sentence being spoken is transcribed as it happens, roughly twice a second, with the settled part solid and the moving end dimmed; the finished transcription of each sentence then replaces it. Can I use the subtitles in OBS? Yes. The overlay can use a chroma-key background for streaming software, is always on top and click-through, and can be kept out of screen captures when you only want it for yourself. Can VoiceMe type into other apps? Yes, but only when you arm it. It types by paste or keystroke into the window you are working in, it is off by default, it needs an explicit arm, and it never types into an app on your block list — with a block list set, it also refuses any window it cannot identify. What does the free version leave out? Nothing but time on the overlay. Listening, transcription, typing and transcripts are unlimited. The floating subtitle window runs for an hour, then pauses for five minutes, then runs again. Pro removes that for €20 once, VAT included. ## Related posts [October 2026: Five Releases, Two New Downloads, One Price List Read](https://royalsoftworks.com/blog/october-2026-releases/) [AssistantGeneral Beta 3: Workspaces, Quick Answers, and an Offline Mode You Can Prove Read](https://royalsoftworks.com/blog/assistantgeneral-beta-3/) [DuplicateDuster 1.4: It Checks Again Before It Deletes Anything Read](https://royalsoftworks.com/blog/duplicateduster-1-4-checks-before-it-deletes/) ## See what we're building Private, on-premise AI and software that runs entirely on your own infrastructure — no cloud, no subscriptions, no data leaving your hardware. [Explore products](https://royalsoftworks.com/products/) [Talk to us](https://royalsoftworks.com/contact/)