Repository navigation
Releases: gomin2013/ai-voice-manager
Release list
AI Voice Manager v1.1.4
v1.1.4 — 2026-10-29
Added
- Check for Updates — in the app menu (under About) and in the About panel: AI Voice Manager asks GitHub for the
latest release and either says you're up to date or shows what's new, with a Download button that opens the DMG.
A daily automatic check can be turned on in About; it is off by default.
Changed
- Privacy Policy — instead of "no update checks" it now says what Check for Updates sends (nothing about you;
GitHub sees your IP address, as with any web request) and when — in all 12 languages. - README — Check for Updates in Features and Privacy.
- Website screenshots retaken from v1.1.3 — Voice Library with 2,000+ voices and its filters, AI Agent
providers, History › Script / Scene — plus a new social preview image.
Installation
Download AIVoiceManager-1.1.4.dmg (and SHA256SUMS.txt), drag AI Voice Manager to Applications (replace
the old copy) and allow the first launch in System Settings → Privacy & Security → Open Anyway (or
xattr -dr com.apple.quarantine "/Applications/AIVoiceManager.app"). Then paste a Gemini API key from
Google AI Studio in AI Providers. Requires an Apple silicon Mac with
macOS 14 or later and an internet connection; an AI Agent provider is optional.
AI Voice Manager v1.1.3
v1.1.3 — 2026-10-27
Fixed
- Two-voice takes generate again — Google's API stopped accepting a setting the app sent for conversations, so
Generate (and a scene preview with two voices) failed with "Unknown parameter 'mode'". One-voice projects were not
affected. - "Generating…" no longer says "(mock — no real API call)": the app does call Gemini.
Changed
- Libraries line up — in Voice, Style and Expression Library, Apply and "✓ in use" share one slot, so ♥ sits in
the same place on every row. - Currency › Match language shows a globe icon instead of an emoji.
- Website: a full demo video — the home page now has a six-minute walkthrough from install to a finished WAV,
narrated in Thai by Gemini TTS, with Thai and English subtitles.
Installation
Download AIVoiceManager-1.1.3.dmg (and SHA256SUMS.txt), drag AI Voice Manager to Applications (replace
the old copy) and allow the first launch in System Settings → Privacy & Security → Open Anyway (or
xattr -dr com.apple.quarantine "/Applications/AIVoiceManager.app"). Then paste a Gemini API key from
Google AI Studio in AI Providers. Requires an Apple silicon Mac with
macOS 14 or later and an internet connection; an AI Agent provider is optional.
AI Voice Manager v1.1.2
v1.1.2 — 2026-10-27
Added
- Language and Pitch filters — Voice Library has Language | Pitch pickers under the tabs again, so 2,000+ voices
narrow down quickly (for example, every low-pitched British English voice). - Demo mode for recordings —
open AIVoiceManager.app --args -demoopens sample projects from a temporary folder,
so screenshots and screen recordings never touch your real projects.
Changed
- "✓ in use" replaces Apply — in Voice Library, Style Library and Expression Library, the item already applied to
the scene shows "✓ in use" where its Apply button was. - Gender shown as ♀ / ♂ — the gender label is the symbol alone (the word is in the tooltip and read by VoiceOver),
so voice rows stay on one line.
Installation
Download AIVoiceManager-1.1.2.dmg (and SHA256SUMS.txt), drag AI Voice Manager to Applications (replace
the old copy) and allow the first launch in System Settings → Privacy & Security → Open Anyway (or
xattr -dr com.apple.quarantine "/Applications/AIVoiceManager.app"). Then paste a Gemini API key from
Google AI Studio in AI Providers. Requires an Apple silicon Mac with
macOS 14 or later and an internet connection; an AI Agent provider is optional.
AI Voice Manager v1.1.1
v1.1.1 — 2026-10-27
Added
- The whole voice library — Voice Library now has Google's entire Gemini TTS library, 2,000+ voices (the same as
Google AI Studio, e.g. Rami), with the 30 portrait voices first. Search by name, description, accent or language;
tabs All · ♂ · ♀ · ♥ favourites · ✓ voices in this project. The list ships with the app (works offline and before a
key) and refreshes from Google about once a week with your key — listing voices is free.
Changed
- Voice names everywhere (timeline, Current Scene, History, previews) show the library's name, and older projects keep
their voices exactly as they were.
Installation
Download AIVoiceManager-1.1.1.dmg (and SHA256SUMS.txt), drag AI Voice Manager to Applications (replace
the old copy) and allow the first launch in System Settings → Privacy & Security → Open Anyway (or
xattr -dr com.apple.quarantine "/Applications/AIVoiceManager.app"). Then paste a Gemini API key from
Google AI Studio in AI Providers. Requires an Apple silicon Mac with
macOS 14 or later and an internet connection; an AI Agent provider is optional.
AI Voice Manager v1.1.0
v1.1.0 — 2026-10-27
Added
- 12 languages — the app now also speaks Japanese, Korean, Vietnamese, Indonesian, Malay, Filipino, Simplified
and Traditional Chinese, Spanish and Russian; search and switch in the Language panel. - AI Agent: five providers — AI Providers has two tabs. Voices: Gemini TTS and Apple's Quick preview.
AI Agent: Apple Foundation Models or Ollama on your Mac, or OpenAI, Gemini (your Google key) or Claude with
your key. The chat shows a step checklist and the model's thinking, renders Markdown with Copy as it streams, and
has Stop; ⚙ opens its settings. Polish and Add pauses use the chosen provider (not shown with Apple). - History › Scene — History has two tabs of 30: Script (the project's steps and pins) and Scene (each
scene's text versions, recorded when you pause typing, with the text it started from). Going back to a Scene row
changes only that scene, so deleted words come back without undoing later timeline edits.
Changed
- History like Hosts Manager — the pin sits in the header, Pinned / Recent sections, a Now badge, right-click
Restore / Rename / Unpin / Pin This Version / Remove, and earlier days show their date. - Prices in Traditional Chinese — the Taiwan dollar shows as NT$ next to US$, never a bare $.
Fixed
- Choosing a language always uses that language's own text (Japanese on a Thai Mac used to stay Thai).
Privacy
- With OpenAI, Gemini or Claude as the AI Agent, your question and the project's script go to that provider with
your key; with Apple or Ollama nothing leaves your Mac. The Privacy Policy and Terms of Use say so.
Installation
Download AIVoiceManager-1.1.0.dmg (and SHA256SUMS.txt), drag AI Voice Manager to Applications (replace
the old copy) and allow the first launch in System Settings → Privacy & Security → Open Anyway (or
xattr -dr com.apple.quarantine "/Applications/AIVoiceManager.app"). Then paste a Gemini API key from
Google AI Studio in AI Providers. Requires an Apple silicon Mac with
macOS 14 or later and an internet connection; an AI Agent provider is optional.
AI Voice Manager v1.0.1
v1.0.1 — 2026-10-24
Changed
- Icons instead of emoji — Getting started ticks, the Ollama card, the Current Scene header, the missing-key
notice and the mute and pin hints in the timeline and Help are drawn with SF Symbols, in the app's own colours. - Website: the whole Gomin family in the footer — every page now links AI Voice Manager too, in the family
order: Gomin Apps · Hosts Manager · Git Exclude Manager · AI Voice Manager · SwiftUI Mac Skills · Profile · GitHub ·
Sponsor.
Installation
Download AIVoiceManager-1.0.1.dmg (and SHA256SUMS.txt), drag AI Voice Manager to Applications (replace
the old copy) and allow the first launch in System Settings → Privacy & Security → Open Anyway (or
xattr -dr com.apple.quarantine "/Applications/AIVoiceManager.app"). Then paste a Gemini API key from
Google AI Studio in AI Providers. Requires an Apple silicon Mac with
macOS 14 or later and an internet connection; Ollama is optional.
AI Voice Manager v1.0.0
v1.0.0 — 2026-10-24
The first release: AI voice-overs written scene by scene, previewed free with the Mac's own voices and generated with
Gemini TTS on your own API key. Every take is kept.
Added
- Projects that keep themselves — Recent Projects on the left; each project is a folder in
~/Library/Application Support/AI Voice Managerand saves on every change. Rename with a double-click or ✎, remove
to the Trash with Undo. Export Project writes one.avmproject.zipwith every take; Open Project… reads it
(or an older.json) on another Mac. - Script Timeline — every line and every pause is a scene: drag to reorder, mute to leave it out of the take, a
voice and a style per line, up to two voices per take (a conversation). - Voice Library — 30 Gemini voices with portraits and descriptions, Male / Female filter, search and ♥ favourites;
12 style presets on the Styles tab. Drag onto a scene or press Apply. - Expression Library — breath, short and long pause, laugh, sigh, gasp and cough between lines, each repeatable up
to ×10. - Listen before you pay — Quick preview reads a line with the Mac's own voices, free and offline; Generate
preview makes one scene with Gemini; Generate (⌘↵) makes the whole take, saved as a WAV with its timeline. - Results — every take with its waveform: play, download the WAV, bring its timeline back, or delete to the Trash
with Undo. - AI Providers — your Gemini API key in the Keychain (Save, Test connection, Remove), Free tier or Billing on;
with billing on the price shows in US dollars and your currency, and every paid take asks first. - AI Agent with Ollama — on your Mac, no login: Polish and Add pauses rewrite one scene (one step in
History), and the chat answers about the whole project as it streams in, with Stop. - Internet alerts — the app says when the connection drops, comes back or slows down, and disables what needs it
while you're offline. - History — your last 30 edits per project and pinned states; going back is a new step, so nothing is lost.
- The Gomin window — Liquid Glass on macOS 26, two tool-window rails, Appearance (16 accents, 8 gradients),
Language & Currency (English and Thai), Help and About; ⌥⌘1…0 open the panels.
Installation
Download AIVoiceManager-1.0.0.dmg (and SHA256SUMS.txt), drag AI Voice Manager to Applications and allow
the first launch in System Settings → Privacy & Security → Open Anyway (or
xattr -dr com.apple.quarantine "/Applications/AIVoiceManager.app"). Then paste a Gemini API key from
Google AI Studio in AI Providers. Requires an Apple silicon Mac with
macOS 14 or later and an internet connection; Ollama is optional.