Desktop app · by KLAWS/FX Licensed · on-device

KlawsFlow

Local AI dictation for Windows. Speak, and it types into any app — in about a second, with nothing leaving your PC.

KlawsFlow turns speech into text wherever your cursor is, and it runs entirely on your machine. No cloud, no account, no audio ever leaving the PC. An optional AI cleanup pass runs on a local model you download with one click — or on your own cloud key, if you choose to bring one. It speaks too: built-in read-aloud voices for English and Greek, and on the Pro tier a voice clone trained from your own dictations reads any text back in your voice, entirely on your GPU.

See plans — from €7.69 See it in action

What it does

Speak, it types

Hold Ctrl+Win, talk, release — your words land as text in whatever app has focus, in about a second.

100% on-device

Speech never leaves your PC. No cloud, no account, works fully offline.

English + Greek, accent-proof

Dual-language pipeline with a dedicated Greek second-opinion model — confident mistranscriptions get vetoed, not typed.

Optional AI cleanup

A built-in local model (or your own API key) fixes punctuation and self-corrections without rewriting what you said.

Your voice, cloned — Pro

Train a voice model from your own dictations and have any text read back in your voice, 23 languages, entirely on your GPU.

Read-aloud voices

Natural built-in voices for English and Greek read any text aloud.

Learns as you dictate

Voice memory keeps tuning recognition to you from the dictations you keep.

Updates itself

Checks on launch, downloads in the background — one restart and you are on the latest version.

See it in action

KlawsFlow Hub home with day streak, words dictated, average WPM, weekly activity and recent dictations — The Hub — your streak, words, and pace at a glance

Dictation history with per-app entries, word counts and WPM — Searchable history of everything you dictated

Voice tab with built-in read-aloud voices, Greek voice, speak-any-text box and constant voice training — Read-aloud voices — and voice cloning that learns from your dictations (Pro)

AI formatting settings with the built-in local AI engine and optional cloud provider — Optional AI cleanup — built-in local model or your own cloud key

Downloads settings listing speech models including Greek Pro and the NVIDIA CUDA engine — Everything downloads inside the app — models, GPU engine, local AI

Buy. Download. Use.

Pick your plan

Pro includes voice cloning; Standard is 30% less without it. Your license key lands in your inbox the moment payment clears.

Run the installer

Installs the app and downloads the speech models once (~2.3 GB).

Paste your key

One key activates one PC. No account, nothing else to set up.

Hold Ctrl+Win and talk

Release, and it types — in any app.

Measured, not promised

≈1s

from release to typed text on short dictations

3×

faster long dictations than our previous engine

100%

of your speech processed on your own PC

audio files uploaded to any cloud, ever

	KlawsFlow	Cloud dictation apps
Where speech is processed	On your PC	On the vendor's servers
Works offline	Fully	No
Audio leaves your machine	Never	Every dictation
Account required	No — just a license key	Yes
AI text cleanup	On-device (optional)	Cloud only

Measured on Windows 11 with an NVIDIA GPU, version 1.6.9. Speed depends on your hardware.

Get KlawsFlow

Monthly (2 PCs)

€10.99/month

1 key · up to 2 PCs

Pro

Everything, incl. voice cloning

Standard

No voice cloning — 30% less