Quantum Talker 1.2.0

Two new ways in

Apple Voice Memos (macOS). Your recordings, listed in the app. Connect the folder
once and every memo sits one click from a clean, punctuated transcript, filed in your
speaking history and ready to copy. Memos you've already done stay marked across
restarts. Prefer zero setup? Dragging memos onto the window still works.

Telegram voice notes. Make your own bot with two messages to @Botfather, paste the
token into Settings, and any voice note or audio file you send that bot comes back as
text in the chat. Record on your phone, read it on your phone. Notes sent while the app
was closed catch up on the next launch.
Two-minute setup guide

This is the one feature that talks to a server, so plainly: it stays off until you
switch it on, it contacts api.telegram.org and nothing else, transcription still runs
entirely on your machine, and your bot token is encrypted with the OS keychain. A chat
with a Telegram bot is ordinary cloud messaging, so give it the same trust you'd give
any Telegram chat.

A transcript viewer. Any transcript opens full-screen and selectable, either whole
or split into sections at sentence boundaries, each with its own copy button. Reachable
from History, from Memos, and from the Scribe tab.

Coming from 1.0.1?

1.1.0 never shipped as a release, so this build also hands you everything that was in
it: dictation anywhere with a hotkey, drag-and-drop transcription of audio files, the
live recording pill with its waveform, a microphone picker, speaking history kept
separate from reading history, and the typing-time-saved dashboard. All of it local, on
NVIDIA Parakeet.

Downloads

Platform	File
macOS (Apple Silicon)	.dmg
Windows (x64)	Setup .exe
Linux (x86-64)	.AppImage

The Mac build isn't code-signed yet, so macOS will hesitate the first time. Right-click
the app, choose Open, and you'll only be asked once.

Models download on first use rather than at install: the voices (~86 MB) on first
launch, the transcription model (~660 MB) the first time you dictate. Everything works
offline after that.

Quantum Talker 1.0.1

Quantum Talker — Open Source & Free AI Talker & MP3 Renderer.

Select text anywhere, hit a shortcut, put your headphones on. 100% local neural voices — no cloud, no account, no telemetry.

Highlights

• 🎙️ Fully offline TTS — Kokoro-82M via ONNX; one ~86 MB model download, then works in flight mode.
• ⚡ Streaming playback — speech starts within ~1.5 s while the rest generates.
• 🕰️ Pitch-true speed control — 0.8×–2× via WSOLA time-stretching: faster speech, same voice.
• 💊 Floating mini player — always-on-top pill with live waveform timeline, ±15 s skips, speed control, drag anywhere. On macOS, click the menu-bar icon for the same player as a dropdown panel.
• 📊 Minutes-saved dashboard — see how much reading time you've reclaimed.
• 📝 Markdown-aware speech — syntax is never pronounced; headings get natural breather pauses.
• 🔊 28 voices (♀ + ♂, US + GB), quality-graded, one-click previews.
• 🎵 MP3 / WAV rendering for instant voiceovers, plus a full reading history.

Downloads

Platform	File
macOS (Apple Silicon)	Quantum Talker-1.0.1-arm64.dmg / .zip
Windows (x64)	Quantum Talker Setup 1.0.1.exe
Linux (x86_64)	Quantum Talker-1.0.1.AppImage
Linux (arm64)	Quantum Talker-1.0.1-arm64.AppImage

macOS: grant the Accessibility permission once so the shortcut can read your selection directly. Windows/Linux builds are unsigned — expect the usual first-run warnings.

Free software, MIT licensed. Read it, verify it, fork it.