Cracker

Usable GUI for text-to-speech services.

Supported text-to-speech

• ESpeak -- https://en.wikipedia.org/wiki/ESpeak
• AWS Polly -- https://aws.amazon.com/polly/
• Google TextToSpeech -- https://cloud.google.com/text-to-speech/

Why all of this

What can you do with it?

I (the author) have difficulty to focus on only reading. The best way to read is to walk and have white noise in the background but that isn't always possible. The other best way is to read both with eyes and ears. When someone reads and I can follow, I don't overly focus on how some words look weird or continue the line even though there is no text remaining. When someone reads what I see, it sets the pace and helps me remember/reference more easily.

Is it only reader?

Text often have footnotes, emoji, and other decorators. These can be visually pleasing but machine needs simple text. Cracker also provide text (de)formatters so to simplify text for Speakers. Examples of deformatters include removal of extensive whitespace, cryptic citation references and wikipedia decorators.

Is this an active project?

Kind of. I update it as I need it. It hasn't been touched in a while because of plenty of problems with (Py)Qt on Linux and increased popularity of Electron (see Pollytron). But, things have changed, and Qt is even better (and Electron is meh). I'm going to update Cracker as needed. Anyone and everyone is welcome to contribute or submit features request.

Installation

It should be enough to install via pip, i.e.

$ pip install cracker

See Usage section below.

Development

Development requires Python 3.14. The project uses uv for dependency management and reproducible environments:

uv sync --all-extras
make test
make format-check
make typecheck
make build

Run make check before opening a pull request to execute Ruff checks, ty type checking, tests, and a package build. See Dependency policy for version constraints and update procedures.

PyQt6 is used for the desktop GUI and multimedia playback. Its binary wheels include the corresponding Qt libraries, so uv sync or pip install cracker installs the required Python and Qt components.

On Ubuntu, the PulseAudio and Qt X11 compatibility libraries may also be required:

sudo apt-get install libpulse-mainloop-glib0 libegl1 libxcb-cursor0 libxkbcommon-x11-0

See macOS development notes for the macOS setup.

Usage

Since this is a GUI on top of AWS Polly it is assumed that one has credentials stored in default directory. This is ~/.aws/credentials on unix based systems.

AWS IAM Identity Center (SSO) profiles can be configured from Config → Speakers. Enter the profile, SSO session, start URL, SSO region, account ID, role, and Polly region, then choose Save and sign in with SSO. Cracker writes the standard AWS shared configuration and starts aws sso login; this sign-in action requires AWS CLI v2.

Cloud backends download audio to Cracker's cache and play it through Qt Multimedia.

Suggested execution command

$ cd cracker
$ python -m cracker.main

Key shortcuts

There's only one global command (read from clipboard). All commands are expected to be called when Cracker is in focus.

Action	Shortcut	Global
Read (clipboard)	Ctr + Shift + Space	Yes
Read (text area)	Ctr + Shift + R	No
Pause / Resume read	Ctr + Space	No
Stop reading	Ctr + Shift + S	No
Reduce (all active)	Ctr + R	No
Reduce (wiki)	Ctr + Shift + W	No
Reduce (citation)	Ctr + Shift + C	No