yap

yap is a free, open-source voice-to-text CLI that converts speech to text at your cursor. It runs on Linux, macOS, and Windows with ~5 MB RAM.

Lightweight voice-to-text for your desktop.

Hold a key. Speak. Text appears at your cursor. Zero idle footprint.

yap is a free, open-source voice-to-text CLI built with Go. It runs as a single static binary on Linux, macOS, and Windows with no runtime dependencies. Audio is transcribed via the Groq Whisper API in ~1–2 seconds, and yap uses just ~5–10 MB of RAM when idle — a fraction of typical desktop dictation software.

terminal
$ yap listen
▶ listening for hotkey... █

Demo GIF placeholder — hold-to-talk workflow

View on GitHub

Features

Hold-to-Talk Workflow

1

Hold

Press hotkey

2

Speak

Record audio

3

Release

Send to API

4

Text

Appears at cursor

Memory

~5-10MB

yap uses just ~5–10 MB RAM when idle with negligible CPU when not recording.

yap typical desktop app ~200MB+

Speed

~1-2s

yap transcribes in ~1–2s after you release the hotkey.

Powered by Groq Whisper API

Portability

yap ships as a single static Go binary with no runtime dependencies.

Linux macOS Windows

Configuration

yap includes a first-class NixOS module and uses standard TOML config.

[transcription]
provider = "groq"

[hotkey]
key = "Super_R"

Trust & Transparency

Zero Telemetry.

Local Environment Variables.

100% Open Source.

No analytics. No phone-home. Your API keys stay in your environment. Audit every line of code.

Bring Your Own Key

$0

recurring software costs.

Use your own Groq API key. Pay only for what you use. No subscriptions, no tiers, no vendor lock-in.

Frequently Asked Questions

What is yap?
yap is a free, open-source command-line tool that converts your voice to text in real time. Hold a hotkey, speak, release — and the transcribed text appears at your cursor. It runs on Linux, macOS, and Windows.
How does yap compare to built-in dictation?
Unlike macOS Dictation or Windows Voice Typing, yap uses only ~5–10 MB of RAM when idle, has no always-on background process, works across all desktop applications, and sends audio to the Groq Whisper API for fast, accurate transcription.
Is yap free?
Yes. yap is MIT-licensed and completely free. You bring your own Groq API key and pay only for the API usage, which is typically fractions of a cent per transcription.
What speech-to-text API does yap use?
yap uses the Groq Whisper API by default. Groq provides extremely fast inference (~1–2 seconds total latency). You can configure the provider in yap’s TOML config file.
Does yap collect telemetry or analytics?
No. yap collects zero telemetry. It never phones home. Your API keys stay in local environment variables, and you can audit every line of the open-source code.