Icespeak

Icespeak is a Python 3.10+ library that makes Icelandic-language speech synthesis easy.

Installation

python3 -m pip install icespeak

To also install the speak command line tool:

python3 -m pip install 'icespeak[cli]'

Local installation

Note: The Azure TTS package currently only supports OpenSSL version 3.0.*

Clone the repository and cd into the folder. Then create and activate a virtual environment:

python3 -m venv venv
source venv/bin/activate

Install minimal set of dependencies to use the library:

python3 -m pip install .

In order to use the CLI interface, speak, install with:

python3 -m pip install '.[cli]'

Alternatively, to install in editable mode with the development dependencies:

python3 -m pip install -e '.[cli]' --group dev

This project uses uv for dependency management, so the above is equivalent to simply running:

uv sync --all-extras

Usage

Before using, set the environment variable for each service you want to use. Most of them take a JSON string, but the OpenAI key is passed as-is:

export ICESPEAK_AWSPOLLY_API_KEY='{"aws_access_key_id":"your-id","aws_secret_access_key":"your-secret","region_name":"us-east-1"}'
export ICESPEAK_AZURE_API_KEY='{"key":"your-azure-key","region":"your-region"}'
export ICESPEAK_OPENAI_API_KEY='your-openai-key'

These can also be placed in a .env file in the working directory.

Output audio files are saved to the directory specified by the ICESPEAK_AUDIO_DIR environment variable. By default Icespeak creates the directory <TEMP DIR>/icespeak where <TEMP DIR> is the temporary directory on your platform, fetched via tempfile.gettempdir().

By default, generated audio files are removed upon a clean exit, but this can be disabled by setting ICESPEAK_AUDIO_CACHE_CLEAN=0.

Text-to-speech

Simple example of TTS, which includes phonetic transcription:

from icespeak import tts_to_file, TTSOptions

text = """\
Þetta er texti fyrir talgervingu. \
Í honum er ýmislegt sem mætti vera hljóðritað, \
t.d. ræður talgerving oft illa við íslenskar skammstafanir, \
tölvupósta eins og ekki.tolvupostur@vefsida.is,
eða prósentur eins og 48,3%, o.fl.\
"""
tts_out = tts_to_file(
    text,
    TTSOptions(
        text_format="text",  # Set to 'ssml' if SSML tags in text should be interpreted
        audio_format="mp3",  # Output audio will be in mp3 format
        voice="Gudrun",  # Azure TTS voice
    ),
    transcribe=True,  # Default is True
)
print(tts_out.file)  # pathlib.Path instance pointing to file on local file system
print(tts_out.text)  # text that was sent to the TTS service (after the phonetic transcription)

Command line tool

Installing with the cli extra provides the speak command:

# Synthesize and play the given text
speak "Góðan daginn og til hamingju með lífið."

# Read the text from a file, pick a voice, and save the audio instead of playing it
speak --file input.txt --voice Gudrun --no-play --out output.mp3 --audio-format mp3

# List the voices available with the API keys you have configured
speak --list-voices

Run speak --help for the full list of options.

License

Icespeak is Copyright © 2025 Miðeind ehf.

This set of programs is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This set of programs is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

The full text of the GNU General Public License v3 is included here and also available here: https://www.gnu.org/licenses/gpl-3.0.html.

If you wish to use this set of programs in ways that are not covered under the GNU GPLv3 license, please contact us at mideind@mideind.is to negotiate a custom license. This applies for instance if you want to include or use this software, in part or in full, in other software that is not licensed under GNU GPLv3 or other compatible licenses.