OpenWhispr

Upstream: OpenWhispr/openwhispr Deckery fork: not yet — used unmodified

Hotkey-activated voice input via OpenAI Whisper. Hold a key, speak, release — the transcribed text is typed into the focused application. Supports both cloud (bring-your-own API key) and fully local Whisper models.

Why it's in Deckery

On a handheld without a physical keyboard, voice input is the most practical text entry method for free-form text (messages, notes, agent instructions). It covers the majority of typing needs without reaching for an external keyboard.

Setup

OpenWhispr can be triggered by a makima binding. Map a button (e.g. the right trackpad click) to the OpenWhispr hotkey in your Steam Deck.toml.

Noise suppression

Outdoor use with the Steam Deck's built-in microphone benefits significantly from a noise suppressor. The steamdeck-dotfiles include an RNNoise PipeWire filter configuration that handles background noise and wind well enough for practical outdoor dictation.

Known limitations

There are open issues in the upstream OpenWhispr project that noticeably affect usability (response latency, occasional missed activations). These are upstream issues — worth checking their issue tracker before filing a Deckery bug.