> press hotkey > speak > release > text injected
> local processing > zero latency > full privacy
$ oiper --backend gpu --model large-v3
> Loading model... done (0.3s)
> GPU acceleration: enabled
> Listening for hotkey...
> Transcription complete: 1.5s
// SYSTEM CAPABILITIES
Global shortcut. Hold to record, release to transcribe. Any application.
Instant text injection into your focused app. Sub-second delivery.
On-device transcription. Zero external calls. Your data stays yours.
Native code with GPU acceleration. Maximum throughput, minimal resources.
Optional online services. Bring your API key. Full control always.
Model selection, backend prefs, LLM transcription for technical content.
// SECURITY PROTOCOL
All speech-to-text processing runs on your hardware
Audio recordings and logs never transmitted externally
Cloud optimization requires explicit user consent and API key
// BENCHMARK RESULTS
> input: 30sec english audio | metric: transcription time
// INITIALIZE
Download OiPer Desktop. Free. Open source. Privacy guaranteed.