v0.9.0: General Dependency Refresh & Python 3.13 Support

General Dependency Upgrade

• Upgraded most dependencies to their latest stable versions to avoid known vulnerabilities in older versions.

Compatibility Note

• live-translation now supports Python 3.13

v0.8.0: Opus Encoding Support & Bug Fixes

Audio Codec Support

• Added Opus codec support for compressed audio streaming.
• --codec option available in both client and server to switch between pcm (raw) and opus (default).
• Backward-compatible with previous raw PCM format.

Client Connection Enforcement

• Server now enforces only one client connection at a time.
• Fixed bug where a second client would be rejected even after the first had disconnected.
• Replaced asyncio.gather() with Python 3.11+ asyncio.TaskGroup for robust task management and error handling.

Compatibility Note

• Python 3.11+ is now required due to usage of asyncio.TaskGroup.

v0.7.0: Client Stability & Testing

• Improved client stability and ensured only one can be connected to server at a time
• Proper testing implemented
• Quality of life development features
• Example client GIFs

v0.6.0: Demo CLI Tool, Dep Cleanup, and Examples

• Cleaned up dependencies by removing requirements.txt and updating dev and example optional dependencies in pyproject.toml

• Added 5 client examples (Node.js, Browser JS, Go, C#, Kotlin/Android)

• Added demo cli tool live-translate-demo to run server-client pair with default settings

• Update to websocketIO to print the uri of the server in red for better visiblity

• Added README badges

• Added useful classifiers to pyproject.toml

• Updated tests to run on a separate port and added coverage to the tests

• Readme update to reflect all changes including instructions on how to integrate the project into non python workflows

• General logging improvements

• Improved examples

v0.5.0: Server-Client General Improvements

• Added large-v3 and large-v3-turbo to the whisper_model options with some extra details about it in the --whisper_model flag help message
• Don't initialize transcription component in the pipeline if in transcribe_only mode
• Allowed server flag to be passed with strings starting with wss in addition to the original ws for more flexibility
• --version option for server and client

v0.4.0: Server-Client Split, Async Support, and WebSocket Audio Streaming

Whats New

• Re-architected the project into a clean server-client model

• The server now handles all audio processing, transcription, and translation

• The client streams microphone audio and receives live results over WebSocket

• Introduced support for both blocking and non-blocking (async) usage patterns

• CLI tools renamed to live-translate-server and live-translate-client

v0.3.2: PyPi Doc Update

PyPi Doc Update

v0.3.1: Cross Platform Update

Updated the program to work on systems that does process "spawn" by default e.g. windows.

v0.2.0: Merge pull request #3 from AbdullahHendy/feature-dep-lock

Locked pyproject.toml dependencies.
Fixed running tests with make test
Minor Docstring updates

v0.1.0: Initial Release

Initial release that supports stdout, file, and websocket outputs. Not all language pairs were tested. Improvements are to be added.