Default Mode
On-Device TranscriptionFast dictation for focused desktop work
Speak once, transcribe instantly, and stay in flow
Hold a global hotkey to record and release to transcribe. OiPer injects text into your active app with local processing by default and optional online services when you choose.
Latency
1.5s / 30 sec audioBackends
Auto, CPU, GPUExperience
Global hotkey workflow designed for uninterrupted thought.
Performance
Native speed with low latency and optional hardware acceleration.
Privacy
On-device default architecture with explicit online opt-in only.
Quick Recording & Transcription
- Hold hotkey to record voice input immediately.
- Release to transcribe and inject text into active apps.
- Choose local or online cleanup after conversion.
Flexible Runtime Control
- Backend selection: auto detection, cpu, or gpu.
- Model choices for speed, quality, and footprint.
- Multiple LLM options for highly technical vocabulary.
Local Processing
Audio and logs remain on-device
OiPer keeps your transcription workflow private by running on your machine, preserving user control and reducing exposure.Online Services (Optional)
Enable with your own API key
Configure base URL, model name, and key when needed, then disable online optimization any time from settings.Settings & Configuration
Deep control, clear interface
Speech model library with simple downloads and quick switching.
Provider settings for API key, custom base URL, and model names.
Optimization mode routing between local and online pathways.
Advanced LLM transcription for domain-heavy terminology.
Performance
Benchmark: 30 sec English audio
OiPer delivers the fastest results while keeping user control and privacy central.