Skip to content

Releases: patrick-1984/handy-tool

Handy Tool 2.4.1

Choose a tag to compare

@patrick-1984 patrick-1984 released this 19 Oct 13:40

Changed

  • Parakeet versions are chosen in Parakeet's settings. The gear on the Parakeet card (or a click on it while in use) opens its versions: the built-in v3 and v2, Parakeet Ultra and Ultra in full precision from Hugging Face (add, download, use), versions you added, and a search for others. A Hugging Face version in use shows on the Parakeet card.

Fixed

  • Parakeet Ultra from Hugging Face is accepted. Ultra exports leave out the audio preprocessor, so Check called them not usable; Handy now supplies its own for the Parakeet TDT 0.6B family (pinned and checked), with word-for-word the same result. Ultra is more accurate than the built-in v3 at the same speed.
  • Old logs are compressed and set aside: at 10 MB the log starts a new file and the old one is compressed (newest 20 kept, up to 30 days).

Handy Tool 2.4.0

Choose a tag to compare

@patrick-1984 patrick-1984 released this 18 Oct 20:00

Added

  • Your own STT server, with speakers (Advanced settings › STT server). Connect any OpenAI-compatible speech-to-text server on this PC or your network; Test lists its models. With "Tell the speakers apart" on, History keeps the take with speakers ("A: …" per turn) next to the plain text, with a Text / Speakers switch, and you choose which one is pasted, each with its own prefix and suffix. Your custom words are sent to the server and corrected in its answer.
  • A second and third model if the first fails (Models › If your dictation model fails). When your dictation model cannot transcribe a take (server down, network dropped, request failed) the next one does; an online model is checked as soon as the take starts. History shows which model transcribed.
  • Parakeet models from Hugging Face (Models › Parakeet from Hugging Face). Search, check and add newer or fine-tuned Parakeet ONNX models; they download file by file (resumable, checksum-checked) and run on this PC.
  • A model's settings from its card: a gear on model cards with settings, and a click on the card of the model in use opens them.

Changed

  • History's "Transcribe again" has a new icon.

Handy Tool 2.3.1

Choose a tag to compare

@patrick-1984 patrick-1984 released this 15 Oct 17:55

Fixed

  • Long recordings play to the end in History. A take longer than 10 minutes is saved in 10-minute pieces, and History's player stopped at the end of the first one (an 11:43 take stopped at 10:09 and could not be moved past it). It now plays and seeks across the whole take. The file on disk is unchanged.
  • What's New lists 2.2.3 and 2.3.0 too. Their entries were missing.

Handy Tool 2.3.0

Choose a tag to compare

@patrick-1984 patrick-1984 released this 15 Oct 11:05

Changed

  • A recording is no longer transcribed twice. Older versions added Handy's recordings folder to the watched folders, so after every take the Translator read the recording again from its file to write handy-….txt - a second full transcription that kept the processor busy. Each take now writes that text file itself, from the text it already has (History › Settings › "Save a text file next to each recording", on by default). The recordings folder is removed from the watched folders once on update; a folder you add yourself stays. Transcribe again in History updates the take's text file, and deleting a take removes it - only when it is the one Handy wrote; a .txt of your own is never replaced or deleted.

Added

  • Drop an audio file to transcribe it. On Files, drag an audio file onto the "Choose an audio file" box.
  • Online models for files. The online models (API, OpenRouter) now work on the Files page whatever your dictation model is; a file's OpenRouter cost is never added to the take you are dictating.

Handy Tool 2.2.3

Choose a tag to compare

@patrick-1984 patrick-1984 released this 15 Oct 09:12

Fixed

  • History player seeks anywhere in long recordings. For takes longer than a minute or two the slider could not be moved to an arbitrary point (it played, but jumping ahead failed or snapped back). The recording is now loaded into memory before it plays, so the whole take is seekable.
  • History player shows the recording's length right away. It showed "0:00 / 0:00" until you pressed play; it now shows the stored length of the take (e.g. "0:00 / 5:19") from the start.

Handy Tool 2.2.2

Choose a tag to compare

@patrick-1984 patrick-1984 released this 13 Oct 20:39

Added

  • Works behind a proxy that asks for a sign-in (Windows). Handy read the proxy's address from
    Windows but could not sign in to it, so on a PC behind such a proxy model downloads and update
    checks failed with "error sending request" while the browser worked. Handy now uses the proxy
    set in Windows the way a browser does - static proxy, bypass list or setup script (PAC) - and
    signs in with your Windows account (Negotiate / NTLM) when the proxy asks. Advanced settings >
    Proxy (new) also offers a manual proxy (address, port, sites reached directly), a user name and
    password for a proxy with its own account (stored encrypted for your Windows account), No
    proxy, and a Test button that shows the route and the sign-in used. Covers model downloads,
    update checks, LLM providers and the online transcription engines; servers on this PC are
    always reached directly.

Fixed

  • Network errors say why. A failed model download is now logged with its cause (it was not
    logged at all), and update-check errors include the underlying reason (DNS, connection, TLS,
    proxy) instead of only "error sending request".

Handy Tool 2.2.1

Choose a tag to compare

@patrick-1984 patrick-1984 released this 13 Oct 01:47

Handy Tool 2.2.1

Fixed

  • Watched folders use the model you picked for files. The Files page has two model choices: Model at the top (for files you add) and Batch model for the watched folders, which defaulted to your dictation model. With FLM (NPU) picked at the top, the watched folders kept transcribing with the dictation model on the graphics card. Batch model's default now follows the model picked for files (and the list says which one); with none picked it is still your dictation model, and a batch model you pick yourself still wins.
  • The watched folders' model is no longer unloaded between the pieces of a file you are transcribing on the Files page.

Handy Tool 2.2.0 - Steady Mic

Choose a tag to compare

@patrick-1984 patrick-1984 released this 13 Oct 00:28

Handy Tool 2.2.0 "Steady Mic"

Recording you can rely on: no false "Starting mic...", short takes transcribed while you speak, FLM (NPU) for files and folders, and every take kept even when transcription fails.

Fixed

  • No more false "Starting mic..." — after a driver update a microphone can take longer to open, and every take then counted as a "cold start"; the quiet pause before your first word was measured as the microphone "warming up". The warm-up is now measured against the room's noise floor, kept per microphone, and the pill waits only after five measured cold starts of that microphone (the shortest fade-in measured). Your words were always recorded; now the pill says so too.
  • FLM (NPU) shows how far the transcription is ("Transcribing N%"), from the second take on.
  • FLM for files and watched folders — pick FLM Whisper on the Files page; it runs on the NPU beside your dictation model. A watched folder on FLM no longer pushes your dictation model out. Online models (API, OpenRouter) are listed but greyed: they work for files only while they are your dictation model.
  • A take whose transcription fails is always kept, with its recording and a History row, however short.

Added

  • Transcribe a recording again from History — new button beside Copy; for a take recorded without a model or one whose transcription failed.

Changed

  • Short takes are transcribed while you speak — after 10 s of speech a clear pause (~0.7 s) sends what you said so far; after 20 s any pause does; 45 s is cut anyway.
  • The text arrives up to 80 ms sooner after the last piece, and the pill goes away sooner.

Full list: CHANGELOG.md

Handy Tool 2.1.4

Choose a tag to compare

@patrick-1984 patrick-1984 released this 07 Oct 11:56

A fix release: Shortcut Keeper stays in front in Remote Desktop, and Handy Tool starts on PCs without Vulkan.

Shortcut Keeper stays in front in Remote Desktop (Windows)

The 2.1.3 log showed why kept shortcuts still reached the remote session. The Windows App puts its own keyboard hook back in front of Shortcut Keeper's a few seconds after you enter a session, and then every 15 to 35 seconds. The 2.1.2 self-heal reacted only after you had typed for 3 seconds unheard, and after mouse-only use it waited up to a minute.

Now, while a session has the keyboard, Shortcut Keeper puts its hook back in front after every second in which it heard no key and no key is held. A key let go while the client's hook was in front no longer makes the next kept shortcut go to the remote.

Still best effort: a shortcut pressed within a second or two of the client taking the front can reach the remote. Press it again, or click outside the session and back in.

Handy Tool starts on PCs without Vulkan (Windows)

On a PC without a Vulkan graphics driver (often an older laptop), Windows refused to start Handy Tool: "The code execution cannot proceed because vulkan-1.dll was not found". Handy Tool now loads the Vulkan library only when Whisper first looks for a graphics card. Without it, Whisper models run on the processor.

A Vulkan driver older than version 1.2 now also counts as "no graphics card", instead of closing Handy Tool when a Whisper model loads.

Install

Windows x64 — Handy.Tool_2.1.4_x64-setup.exe below. The installer is not code-signed with Authenticode, so SmartScreen will warn: More info → Run anyway. Existing installs update in place: click the version in the sidebar → Update now, or let the nightly silent update do it. (winget still offers 2.0.3; Handy updates itself from there.)

Linux — rpm, deb or AppImage. These are built but have not been tested on real hardware.

macOS — the disk images are built, but the app does not work on macOS yet.

Handy Tool 2.1.3

Choose a tag to compare

@patrick-1984 patrick-1984 released this 07 Oct 00:37

A diagnostic release for Shortcut Keeper.

Shortcut Keeper says why it did not keep a shortcut (Windows)

When you press a kept shortcut in a Remote Desktop session and Shortcut Keeper lets it through instead of running Handy, the log now names the shortcut and the reason:

  • the hook not ready yet (just entered the session),
  • the key or a modifier already down when the hook went in,
  • a Windows key held,
  • the keyboard focus outside the session window, or a full queue.

Only a kept shortcut's whole key combination, pressed with the physical keys, is ever noted. Ordinary typing, AltGr characters (ó, ł, @…) and keys sent by software (password managers, macro tools) never are.

A shortcut that reached the remote with no such line got past Handy's hook entirely, because the Remote Desktop client's own hook was in front. This tells the two apart the next time a shortcut such as Ctrl+Alt+P lands in the remote session — where, in Codex or Claude Code, it recalls your previous prompt — instead of starting a recording.

Tip: a Transcribe & Submit shortcut without a letter (for example Ctrl+Shift+F2) does no harm if it ever reaches the remote.

Install

Windows x64 — Handy.Tool_2.1.3_x64-setup.exe below. The installer is not code-signed with Authenticode, so SmartScreen will warn: More info → Run anyway. Existing installs update in place: click the version in the sidebar → Update now, or let the nightly silent update do it. (winget still offers 2.0.3; Handy updates itself from there.)

Linux — rpm, deb or AppImage. These are built but have not been tested on real hardware.

macOS — the disk images are built, but the app does not work on macOS yet.