Double-tap or hold to activate
Default: double-tap — doesn't conflict with ⌥ special characters like @. Alternatively push-to-talk (hold). Freely assignable shortcut, e.g. ⌥ Space. Esc discards the recording.
Dictate.
Locally.
Done.
Hold the hotkey → speak → finished text lands in the active app.
No cloud service, no subscription, no third-party app. 100 % offline.
macOS 13+
Apple Silicon
AGPL v3
No subscription
Three promises. No exceptions.
100% local
Your audio never leaves your Mac. whisper.cpp runs directly on your hardware — no cloud, no server, no outbound calls.
No subscription, no account
Build once, use forever. No sign-up, no subscription, no tracking. The tool is yours.
Open source — AGPL v3
Full source code on Codeberg. An independent implementation built on whisper.cpp and Ollama.
What Voxista does
A native macOS menu bar app — lean, precise, no bloat.
Enhancement modes
Raw · Clean up · As email · Translate to English — plus custom actions in the editor. Optionally use a cloud provider (Langdock / OpenAI-compatible) instead of Ollama locally.
Warm mode
"Keep model in memory" starts whisper-server in the background — transcription begins instantly, no model load time on first dictation.
Live preview in the HUD
Experimental: a running transcript appears every ~2.5 s while you speak — requires warm mode.
Dictionary & snippets
Automatic word replacement (dictionary.txt), a proper noun list (vocabulary.txt) for the LLM, and text snippets via trigger word — dictate "Signature" and the snippet is inserted.
History
Recent transcriptions accessible in the window and directly from the menu bar menu. Search, pin, delete individually, drag & drop out.
Meeting detection
If another app is using the microphone (Teams, Zoom & co.), Voxista offers to transcribe your side of the conversation — your audio lands as a take.
Recording HUD with waveform
A floating pill shows live status: real-time waveform, word counter "✓ 128 words inserted", 10-second countdown on silence.
Automation & CLI
The same binary as a command-line tool — for Shortcuts, the URL scheme voxista://, and an optional webhook (e.g. n8n).
How it works
Four steps from hotkey to text in the active app — entirely on your Mac.
Record
Hotkey & microphone
Hold or double-tap ⌥ — Voxista records at 16 kHz WAV. The menu bar icon pulses red. Esc cancels.
Transcribe
whisper.cpp locally
Large-v3-Turbo runs entirely on your hardware. Long recordings are split at silence points and transcribed with progress updates.
Optional
LLM enhancement
Ollama (qwen2.5) cleans up, formats, or translates the raw text — depending on the selected mode. Dictionary and vocabulary are applied.
Output
⌘V into the active app
Text is pasted via clipboard + ⌘V exactly where the cursor is. Optionally followed by an automatic Return.
Scripting & integration
Voxista as a building block in your automation stack — Shortcuts, shell scripts, webhooks.
# ./build.sh --install sets up the command automatically voxista --transcribe interview.mp3 # audio → text (local) voxista --enhance "um the meeting is tomorrow" --mode clean voxista --list-modes # list available modes
URL scheme
voxista:// — trigger actions directly from other apps or the browser.
Webhook
Each completed dictation is optionally sent to a webhook (e.g. n8n) — secured with a Bearer token stored in the macOS Keychain.
Shortcuts app
"Run shell script" action — uses the same engine and settings as the running app.
Getting started
The fastest path is the prebuilt download. If you prefer, build from source instead — both options are described below.
Option A — Download (recommended)
# 1. Download & unzip, drag Voxista.app to /Applications: # → codeberg.org/kleinschmidt/Voxista/releases/latest # 2. First launch: right-click the app → "Open" → "Open" # (Voxista is self-signed, not notarized — one-time prompt only) # Alternatively, in Terminal: xattr -dr com.apple.quarantine /Applications/Voxista.app # 3. The Whisper model is downloaded automatically on first launch (guided setup)
whisper is already bundled in the download — Homebrew is not required. The app notifies you about updates (daily version check, no auto-install).
Option B — Build from source · 1 — Prerequisites
# Homebrew (brew.sh) and Swift Command Line Tools are required # Full Xcode is not needed # Speech recognition brew install whisper-cpp mkdir -p ~/whisper-models curl -L -o ~/whisper-models/ggml-large-v3-turbo-q5_0.bin \ https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-large-v3-turbo-q5_0.bin # Optional LLM enhancement (local) brew install --cask ollama-app ollama pull qwen2.5:7b
2 — Build & install
git clone https://codeberg.org/kleinschmidt/Voxista cd Voxista ./setup-signing.sh # once: create a stable signing certificate ./make-icon.sh # generate app icon ./build.sh --install # build and install to /Applications # The built app runs without Homebrew (whisper is bundled)
First launch: In System Settings → Privacy & Security allow Voxista under Accessibility (global hotkey + paste) and grant Microphone access when prompted on first dictation.
Sharing as a ZIP: Voxista is self-signed (ad-hoc), not notarized. Recipients need to lift the Gatekeeper block once:
xattr -dr com.apple.quarantine /Applications/Voxista.app
Or: right-click the app → "Open" → "Open".
Changelog
What each version brought — newest first.
1.17
June 14, 2026
- Third-party license texts are now bundled and linked from the About tab (plus a warranty disclaimer)
- Contact address updated to hello@heinemann.media
1.16
June 12, 2026
- Terminal (CLI), error messages and first-run templates are now fully bilingual
- AI prompts and meeting speaker labels follow the app language
- Stability: hardened context capture against rare crashes
1.15
June 12, 2026
- Voxista now speaks English: the interface defaults to English and can be switched to German (Settings → General)
- All menus, panels and dialogs are fully bilingual
- Built-in AI prompts and meeting speaker labels follow the app language
1.14
June 12, 2026
- Voxista is now available for download: prebuilt releases on Codeberg (signed ZIP)
- Update notice: the app checks for new versions daily — one click in the menu takes you to the download (no auto-install)
- About tab: “Check for updates” button
1.13
June 12, 2026
- Renamed: Voxify becomes Voxista — logo and features stay the same
- All your data (settings, dictionary, takes, history, API keys) migrates automatically
- The CLI is now “voxista”; URL scheme voxista:// (voxify:// remains as an alias)
1.12
June 12, 2026
- Meeting recording (phase 2): now captures the other side too — stereo take (you on the left, everyone else on the right)
- Speaker transcript: “Me:” / “Others:” interleaved by time, right from the meeting panel (also via CLI)
- One-click AI minutes: topics, decisions, action items, open questions (new “Meeting minutes” mode)
1.11
June 11, 2026
- Meeting detection (phase 1): when another app holds the microphone and Teams/Zoom & co. are running, Voxista offers to transcribe your side — your audio lands as a take
- New working logo: Voxista brand mark as app icon, in the About tab and onboarding (vectorized in Branding/)
1.10
June 11, 2026
- Long recordings (> 5 min) are split at silence points and transcribed with progress ("Part 2 of 5") — much more reliable than processing in one go
- HUD position choosable: bottom or top center
- Webhook can be secured with an auth token (Bearer, Keychain); donate button in the About tab
1.9
June 11, 2026
- Dictation language (German / English / Auto) switchable directly from the menu
- Text snippets: dictate a trigger word (e.g. "Signature") → snippet is inserted
- Automation: CLI with
--json/--stats/--export-history, URL schemevoxista://, optional webhook (n8n) - Polish: confirmation dialogs before deleting, auto-save on exit, visible errors
- HUD word count: "✓ 128 words inserted"
1.8
June 11, 2026
- Takes list: every recording is kept (timestamp), with player, transcribe, export and individual delete
- Silence auto-stop: after extended silence the HUD counts down 10 s — speaking again cancels it (limit configurable)
1.7
June 11, 2026
- Live preview in the HUD (experimental): running transcript every ~2.5 s — requires warm mode
- Freely assignable hotkey (e.g. ⌥ Space) in addition to right modifier keys
- Esc cancels an active recording; "✓ Inserted" confirmation after dictation
- Menu bar icon reflects state (pulsing red while recording)
- History: search, pin, delete individually, drag & drop entries out
- Actions editor: "Try it" tests the prompt with sample text
- Guided first launch (microphone, Accessibility, model) instead of silent failure
- Recording tab with live waveform; system mute applies there too
1.6
June 11, 2026
- Warm mode "keep model in memory": whisper-server keeps the model loaded — transcription starts immediately
- Dictation history accessible directly from the menu bar menu
- New recording HUD with live waveform
- Bug fixes: download status check, clipboard protection
- Performance improvements: file caches, faster settings
- First unit tests
1.5
June 10, 2026
- File transcription via drag & drop
- Permissions overview with live status
- Microphone selection
- Waveform player
- Visible enhancement errors
1.4
June 10, 2026
- Settings with sidebar
- Cloud provider for AI enhancement (Langdock/custom, keys in Keychain)
- Editable preview before inserting
- CLI mode for Shortcuts
1.2 / 1.3
June 10, 2026
- Double-tap activation
- Settings window
- Custom models
- App-specific modes
- Context from selected text
- History
- Level meter
1.1
June 10, 2026
- Dictionary & vocabulary
- Hybrid hotkey
- VAD (Voice Activity Detection)
- Model management
- Sound feedback
1.0
June 10, 2026
- Initial release
- Push-to-talk dictation
- 4 modes (Raw, Clean up, As email, Translate to English)
- Local Whisper + Ollama
Voxista stays free
Voxista is open source (AGPL v3) and comes without a subscription, without an account, without hidden costs. If it helps you day to day, consider supporting development via Liberapay — completely voluntary and anonymous.
External link — Liberapay has its own privacy policy.