Local speech-to-text with word-level timestamps, text-to-speech, live dictation, and language ID. Install the engine and models deliberately, see progress on stderr, and keep your audio on your machine. CoreML/ANE on Apple Silicon; ONNX on Linux and Windows. No Python. No ffmpeg. No surprise downloads.
$ bun add -g @drakulavich/kesha-voice-kit
Capabilities
Four pipelines, one binary. Pick what you need; the rest stays out of your way.
25 languages via NVIDIA Parakeet TDT. ~19× faster than Whisper on Apple Silicon, ~2.5× on CPU. Auto language detection, optional VAD for long audio, word-level timestamps, plain text, or LLM-friendly TOON output.
Drop a meeting recording in, get back timestamped segments tagged with a stable
speaker ID. Powered by FluidAudio’s Sortformer model running on the Apple Neural
Engine via a Swift sidecar — ~4.6× real-time on a 71-minute Zoom call. v1.19
gives the cold ANE warmup extra headroom so the first --speakers call
on a fresh process no longer hits a false timeout.
Kokoro runs through CoreML/ANE on Apple Silicon and ONNX on Linux and Windows;
Vosk-TTS covers Russian, and macOS system voices need no model download. macOS
auto-routes from the text language; Linux and Windows use --lang or
--voice. Kokoro (ANE or ONNX) and Vosk support SSML
<prosody rate> and <break>;
<emphasis> is Vosk-only. macOS system voices reject
--ssml.
Synthesize straight into the format Telegram, WhatsApp, Signal and Discord expect for inline voice messages — ~24× smaller than WAV. Pick a bitrate; the engine handles the OggOpus container, no ffmpeg required.
--json --timestamps adds a words array to each segment:
file-relative predictions on a 0.08-second grid. Each word end is predicted
independently, so consecutive spans may overlap and an end is not necessarily the
next word’s start. The field is absent when the decoder has no usable words (for
example after --itn), not an empty array.
107 languages from audio via SpeechBrain ECAPA-TDNN. JSON and TOON also expose
textLanguage separately from spoken audioLanguage:
Apple’s recognizer on macOS and bundled tinyld elsewhere. Their confidence scales
are not comparable.
Silero VAD v5 trims silence on long, sparse recordings. Install it explicitly
with kesha install --vad; Kesha uses it automatically past 120
seconds once it is available.
--stdin-loopOne TTS process, many requests. Models stay in memory across calls so each synthesis hits ~10 ms cold-start instead of ~800 ms. Ideal for chatbots, live agents, and long agentic loops.
On darwin-arm64, kesha record --live prints the microphone transcript
to stdout. Install VAD explicitly, then opt into silence stopping with
--auto-stop: 1,000 ms silence after 250 ms speech by default. Live
recording has a 120-second hard stop by default; pass --max-seconds
to override it. Ctrl-C prints the transcript; the recovery WAV is deleted only
after uninterrupted successful stdout delivery, and is kept if delivery fails or
Ctrl-C or SIGTERM ends the session.
Opt-in local-only command events with paths and user-supplied text redacted.
kesha doctor surfaces status; kesha logs manages them;
kesha support-bundle bundles them for support. Mode
retain-on-failure keeps logs only when a command actually fails.
Install TTS only when you need it. Kokoro powers English and multilingual voices
through CoreML/ANE on Apple Silicon and ONNX on Linux/Windows; Russian uses
Vosk-TTS. Chinese, Japanese, and Hindi are darwin-arm64 only;
Chinese accepts Han text, while Hindi requires romanized Latin input and Japanese
requires romaji (Devanagari and kana/kanji are rejected). Use --lang
on Linux and Windows rather than relying on macOS auto-routing.
Every failure path prints a stable, machine-readable error [CODE]: …
line. kesha-engine --error-codes-json dumps the whole taxonomy; a
docs drift gate keeps docs/errors.md honest. Agents and CI can branch
on why something failed instead of regexing log soup.
A single Rust binary for macOS arm64, Linux x64, and Windows x64. FluidAudio
(CoreML) on Apple Silicon via dedicated sidecars (diarization, TTS, language
detection); ort (ONNX) elsewhere. Symphonia decodes WAV, MP3, OGG/Opus, FLAC,
AAC, M4A; pure-Rust opus + flacenc handle encode
— royalty-free codecs, no C dep, no patent exposure. Linux .deb and
.rpm packages signed with Sigstore.
No ffmpeg. No Python. No native Node addons.
Plug straight into your LLM agent as a voice-processing skill. Give Claude (or any agent) ears and a voice in one install.
Hear it
Real output from kesha say — Kokoro-82M for English, Vosk-TTS for Russian.
SSML <prosody rate> renders the same text at different speeds without
resampling artifacts. Kokoro and Vosk honor <prosody> and
<break>; <emphasis> is Vosk-only.
“Kesha turns voice into text in a single command — no Python, no ffmpeg.”
kesha say --voice en-am_michael "…" --out demo.flac
Same text, x-fast prosody. Tempo only — pitch and timbre unchanged.
kesha say --voice en-am_michael --ssml '<speak><prosody rate="x-fast">…'
«Кеша превращает голос в текст одной командой — без зависимостей.»
kesha say --voice ru-vosk-m02 "…" --out demo.flac
Same text, slow prosody. Useful for accessibility and dictation.
kesha say --voice ru-vosk-m02 --ssml '<speak><prosody rate="slow">…'
Performance
Compared against Whisper large-v3-turbo across Russian and English
clips — every engine left to auto-detect language. CoreML on M-series lands at roughly
19× real-time; the ONNX path on x86 still beats Whisper by ~2.5×.
Quick start
Requires Bun ≥ 1.3 on macOS arm64, Linux x64, or Windows x64. Engine and models download only when you ask them to.
Skip if you already have it.
macOS / Linux
$ curl -fsSL https://bun.sh/install | bashWindows PowerShell
> powershell -c "irm bun.sh/install.ps1 | iex"
Preview the size, then explicitly download the engine and base STT models.
Progress and recovery hints go to stderr; Kesha never downloads them while you
transcribe or speak. Add --tts or --vad only when those
workflows need them.
$ bun add -g @drakulavich/kesha-voice-kit
$ kesha install --plan && kesha install
$ kesha install --vad # only for live auto-stop / VADPipe-friendly: stdout is the transcript; stderr carries progress and errors.
$ kesha audio.ogg
Свободу попугаям! Свободу!# Plain transcript
$ kesha audio.ogg
# Text + language/confidence
$ kesha --format transcript audio.ogg
# Full JSON with lang fields
$ kesha --format json audio.ogg
# Compact, LLM-friendly TOON
$ kesha --toon audio.ogg
# Long / silence-heavy audio (after: kesha install --vad)
$ kesha --vad lecture.m4a
# Segments plus word-level timestamps
$ kesha --json --timestamps interview.ogg
# Meeting transcript with speaker IDs (darwin-arm64)
$ kesha --json --vad --speakers meeting.m4a
# Warn if detected language differs
$ kesha --lang en interview.wav
# Live dictation: darwin-arm64 + explicit VAD install
$ kesha install --vad && kesha record --live --auto-stop# Opt-in TTS pack: English; preview sizes first
$ kesha install --tts
# macOS auto-routes from text language
$ kesha say "Hello, world" > hello.wav
$ kesha say "Привет, мир" > privet.wav
# Linux / Windows: state the language explicitly
$ kesha say --lang ru "Привет, мир" > privet.wav
# OGG/Opus voice notes (drop straight into Telegram)
$ kesha say --format ogg-opus "Уже бегу!" --out voice.ogg
# Kokoro / Vosk only: variable-speed playback via SSML <prosody rate>
$ kesha say --voice en-am_michael --ssml \
'<speak><prosody rate="x-fast">Read this fast.</prosody></speak>' \
--out fast.wav
# Use a macOS system voice
$ kesha say --voice Milena "Чао!"Want to hear what these voices sound like? Play the four samples ↗
import { transcribe, downloadModel } from "@drakulavich/kesha-voice-kit/core";
// One-time install
await downloadModel();
// Transcribe a file
const text = await transcribe("audio.ogg");
// With detection metadata
const result = await transcribe("audio.ogg", { format: "json" });
console.log(result.language, result.confidence);# Multiple files — headers per file, like `head`
$ kesha freedom.ogg tahiti.ogg
=== freedom.ogg ===
Свободу попугаям! Свободу!
=== tahiti.ogg ===
Таити, Таити! Не были мы ни в какой Таити! Нас и тут неплохо кормят.
# Pipe transcripts into your favorite LLM
$ kesha --toon meeting.m4a | llm "Summarize action items"
# Status of installed engine + models
$ kesha status
# Diagnose env, log status, and CoreML caches
$ kesha doctor
# Bundle logs + manifests for a bug report
$ kesha support-bundleWhat's inside
Each model is best-in-class for its task and runs through
kesha-engine — Rust, with CoreML sidecars on Apple Silicon and ONNX
everywhere else. Nine multilingual voices ship in one bundle as of v1.22.
Audio decoding via Symphonia — WAV, MP3, OGG/Opus, FLAC, AAC, M4A.
Languages
Need detection only? VoxLingua107 covers 107 languages.
Integrations
Give your LLM agent ears. Voice-processing skill, pre-wired.
Setup guide →
macOS launcher commands: transcribe selected audio, speak the clipboard. From the
raycast/ directory.
Import transcribe() and downloadModel() directly into any
Bun/Node project.
Behind a corporate proxy or fully offline? Point Kesha at your model mirror.
Mirror docs →Open source, MIT licensed, and improving every week. Star the repo or jump straight to the install.