OiPer converts voice to text through a single hotkey — running entirely on your machine at speeds that cloud services cannot match.
Designed around a single principle: get out of your way. Every feature serves the goal of frictionless, private voice transcription.
One key combo to rule them all. Works in every application without configuration.
Text appears directly in your active window. No clipboard, no paste required.
Whisper-based transcription runs fully on-device. Multiple model sizes available.
Auto-detects your hardware. Uses GPU when available for maximum throughput.
Optional LLM-based text polishing — locally or via your own API key.
Backend selection, model choice, language setting. Configure what matters to you.
Every design decision in OiPer starts with a single question: does this require sending data anywhere? If yes, it does not ship as a default.
Speech-to-text runs on-device. No exceptions, no fallbacks to cloud.
All logs and audio stay on your machine. We have no access.
Online features require your own API key and explicit consent.
No analytics, no usage data, no behavioral tracking of any kind.
OiPer is free. No subscriptions, no accounts required. Download and start transcribing in under a minute.
“The fastest local transcription available. 1.5s for 30s of audio, with your data never leaving your device.”
Performance Benchmark, 2026