SpeakoFlow
SpeakoFlow is an open-source desktop voice assistant under MIT that builds on Handy's dictation core. It dictates into any app offline with whisper.cpp or Parakeet, and adds a "Hey Flow" trigger phrase that turns speech into a finished reply or draft, a 795 MB local cleanup model for English, screen-aware answers through a model you choose, and spoken replies with Kokoro. It runs on Windows, macOS and Linux with no account and no telemetry.
What does SpeakoFlow do?
SpeakoFlow calls itself "a free, local voice assistant for your desktop. Dictation, writing, and an AI assistant, all by voice." Dictation works by hotkey into any app, live or on release, and the README says "Transcription runs on your GPU or CPU with whisper.cpp or Parakeet, fully offline." Its assistant layer starts with a trigger phrase, "Hey Flow" by default and renameable, that turns what you say into a finished reply or a draft pasted at the cursor.
The assistant can answer from an offline llama.cpp model built in, from Ollama or LM Studio, or from any OpenAI-compatible cloud key. Screen vision captures the screen only on request and sends it only to the chosen provider, keeping a small thumbnail locally. Replies stream in a floating panel and can be read aloud with Kokoro locally or with OpenAI-compatible, ElevenLabs and Azure voices. A cleanup model called SpeakoFlow Mini, a 795 MB download, strips filler and applies spoken edits such as "new paragraph" and "scratch that", in English only. Existing .gguf and Whisper .bin files are used in place.
Key facts
- Licence: MIT, with two copyright lines, Abhishek Barali (2026) and CJ Pais (2025), the author of Handy; the README says the Handy core is "used under the MIT licence."
- Transcription offline with whisper.cpp or Parakeet on GPU or CPU; Silero VAD for silence; Kokoro for local text to speech.
- Assistant providers: a built-in offline llama.cpp model, Ollama or LM Studio, or any OpenAI-compatible cloud key.
- SpeakoFlow Mini cleanup model: 795 MB, English only, its licence not stated in the README.
- Screen vision only on request, sent only to the chosen provider, with a small thumbnail kept locally.
- Installers: Windows .exe (unsigned), macOS .dmg (unsigned, needs the quarantine attribute removed), Linux AUR package, .deb built on Ubuntu 24.04, and AppImage, for x86_64 and ARM64; from source with Rust and Bun.
- "There is no telemetry and no account."
- Latest tagged release when read: v1.4.0 on 2026-08-30; 245 stars on 2026-09-17.
What does it replace, and where does it fall short?
SpeakoFlow replaces Wispr Flow and Superwhisper for someone who wants the dictation those apps sell plus a voice assistant on the same hotkey, on all three desktops, with nothing sent anywhere unless they choose a cloud model. It is Handy with a cleanup model and an assistant on top, and it appears beside Handy on the open-source alternatives to Wispr Flow page.
Where it falls short: the installers are unsigned, so Windows and macOS warn on first launch. The cleanup model is English only. There is no .rpm package yet, the Intel Mac build is CPU only, and on native GNOME Wayland the overlay cannot stay on top. Hardware minimums are not stated. The project is young, at 245 stars and a first release in 2026.
How does SpeakoFlow run?
Download the installer for Windows, macOS or Linux from GitHub Releases, or install the AUR package on Arch. On macOS, because the app is not Apple-signed, the README gives the xattr command to clear the quarantine flag. Models download in chunks on first use, or the app reuses .gguf and Whisper .bin files you already have. Build from source with Rust and Bun using bun run tauri dev. There is no paid tier and no hosted service.
Who is SpeakoFlow for?
Someone who wants Handy's offline dictation with a cleanup pass and a voice assistant, on Windows, macOS or Linux, and does not mind unsigned installers. Someone who wants a mature, signed app should pick Handy or VoiceInk.
What limits does the README state?
From the README: no .rpm yet because the packaging does not bundle the speech engine correctly; macOS blocks the first launch because the app is not Apple-signed; SpeakoFlow Mini handles English only for now; the overlay cannot stay on top under native GNOME Wayland and uses XWayland instead; the Intel Mac build is CPU only.
Questions people ask
Is SpeakoFlow open source?
Yes. The code is MIT, an OSI approved licence, and the README credits Handy's dictation core, also MIT, as its foundation. The SpeakoFlow Mini cleanup model and the third-party engines it bundles, whisper.cpp, Parakeet, Silero VAD, llama.cpp and Kokoro, carry their own licences, which the README does not list.
Does SpeakoFlow work offline?
Yes for dictation, cleanup and a basic assistant: whisper.cpp or Parakeet transcribe locally, SpeakoFlow Mini cleans up locally, a built-in llama.cpp model answers locally, and Kokoro speaks locally. Cloud providers are optional, and screen captures go only to the provider you chose, on request.
Which platforms does SpeakoFlow support?
Windows, macOS on both Apple Silicon and Intel, and Linux on x86_64 and ARM64, with a .deb, an AppImage and an AUR package. The Windows and macOS installers are unsigned, so both systems warn on first launch, and the README gives the workaround.
How does SpeakoFlow compare with Wispr Flow?
SpeakoFlow does the dictation and cleanup Wispr Flow sells, offline, plus an assistant that answers questions and drafts text by voice, on three desktops, for free with no account. Wispr Flow is a signed, supported product with phone apps. The open-source alternatives to Wispr Flow page compares them.
How is SpeakoFlow related to Handy?
It is built on Handy's dictation core, which the README says is used under the MIT licence, and Handy's author holds one of the two copyright lines in SpeakoFlow's LICENSE. Handy stays a one-job dictation tool; SpeakoFlow adds the cleanup model, the trigger phrase, screen vision and spoken replies.
Sources
- SpeakoFlow README and MIT LICENSE: github.com/AbhishekBarali/SpeakoFlow, read 2026-09-17.
Compared with the others
On the open-source alternatives to Wispr Flow page, SpeakoFlow is the pick for handy plus a cleanup model and an assistant. MIT, built on Handy's dictation core, with a trigger phrase, a local English cleanup model, screen-aware answers and Kokoro speech, on all three desktops.
Also on that page: Handy for all three desktops, nothing else running, VoiceInk for a mac, with modes per application, FluidVoice for a mac with the newest local models.
On the open-source alternatives to Superwhisper page, SpeakoFlow is the pick for handy plus a cleanup model and an assistant. MIT, built on Handy's dictation core, with a trigger phrase, a local English cleanup model, screen-aware answers and Kokoro speech, on all three desktops.
Also on that page: VoiceInk for a mac, with modes per application, Handy for all three desktops, nothing else running, FluidVoice for a mac with the newest local models.
More voice dictation apps
A free, offline push-to-talk dictation app for macOS, Windows and Linux that runs Whisper or NVIDIA Parakeet on your own machine and pastes the text at your cursor, under MIT.
A GPL-3.0 macOS dictation app that runs Nemotron, Parakeet, Cohere Transcribe, Apple Speech or Whisper on the device, with command and write modes, per-app prompts, and an optional closed local enhancement model.
A dictation and meeting-notes app under MIT that runs Whisper or Parakeet locally or cloud models with your own key, detects Zoom, Teams and FaceTime calls, and labels speakers on device.
A macOS dictation app under GPL-3.0 whose README names Superwhisper and Wispr Flow as the tools it replaces, with local models, per-app modes, a personal dictionary and an AI assistant mode.
A speech-to-text app inside the Epicenter monorepo under AGPL-3.0 that records, transcribes with a provider you choose, cloud or local, optionally polishes the text, and pastes it at the cursor.
An AGPL-3.0 voice typing app for Windows that ships seven local speech models including Nemotron 3.5 ASR and three Qwen3-ASR sizes, with an editable AI cleanup step through Ollama or any OpenAI-compatible endpoint.
Added September 17, 2026. Every claim above comes from the project's README, LICENSE or model card, read on September 17, 2026, or from the GitHub API on the date shown in the panel. Found an error? Write to reveneau@licheo.com and it is fixed in the next weekly pass. Repository: github.com/AbhishekBarali/SpeakoFlow.
Running one of these inside your own environment, with your own data and your own security rules, is the kind of work Reveneau does. Read how a forward deployed engagement works.