What's Changed

OpenAI diarization and streaming support:

• Added support for OpenAI's gpt-4o-transcribe-diarize model, enabling diarization (speaker labels and timestamps) and streaming output in the transcription pipeline (pkg/manager.go, README.md).
• Modified the OpenAI transcription flow to emit streaming segments via a SegmentWriter interface, handling both diarized and non-diarized streaming events (pkg/manager.go).

Refactoring for streaming and extensibility:

• Refactored the transcription and translation methods for Whisper, OpenAI, and ElevenLabs to accept a SegmentWriter, enabling real-time segment emission for all providers and improving code consistency (pkg/manager.go).

Error handling improvements:

• Standardized error handling by replacing custom errors with httpresponse errors for not found and bad parameter cases, improving error reporting to clients (pkg/manager.go, pkg/httphandler/httphandler.go).

Full Changelog: v0.0.38...v0.0.39

Downloading the CLI

Download the CLI for Linux and MacOS here. This does not contain the server. To check the integrity of the binary, and make it executable:

# Download and check against the SHA256 hash on the releases page (MacOS)
wget -O gowhisper -q https://github.com/mutablelogic/go-whisper/releases/download/v0.0.39/gowhisper-darwin-arm64
sha256sum gowhisper
chmod +x gowhisper

# For MacOS, remove quarantine attribute
xattr -d com.apple.quarantine gowhisper

# Run the executable
./gowhisper --help

What's Changed

• Added windows support for libwhisper bindings by @djthorpe in #92
• Added windows CLI to the build process
• Added multipart streaming for transcribe and translate from the client-side to reduce memory footprint
• Refactored tasks and include otel spans on whisper segments

Full Changelog: v0.0.37...v0.0.38

Downloading the CLI

Download the CLI for Linux and MacOS here. This does not contain the server. To check the integrity of the binary, and make it executable:

# Download and check against the SHA256 hash on the releases page (MacOS)
wget -O gowhisper -q https://github.com/mutablelogic/go-whisper/releases/download/v0.0.38/gowhisper-darwin-arm64
sha256sum gowhisper
chmod +x gowhisper

# For MacOS, remove quarantine attribute
xattr -d com.apple.quarantine gowhisper

# Run the executable
./gowhisper --help

What's Changed

Dependency Updates:
Updated github.com/mutablelogic/go-client from v1.3.3 to v1.3.4 and github.com/mutablelogic/go-server from v1.5.17 to v1.5.18 in go.mod for improved stability and bug fixes. Also replaced github.com/djthorpe/go-pg with github.com/mutablelogic/go-pg v1.1.1 as an indirect dependency. [1] [2] [3]
Updated the third_party/whisper.cpp submodule to a newer commit for latest upstream improvements.

HTTP Client Improvements:
Added the ability to pass custom request options via a new WithRequestOpts function and reqOpts field in the opt struct in pkg/httpclient/opts.go, allowing more flexible HTTP requests. [1] [2]
Refactored request option handling in both Transcribe and Translate methods to utilize the new reqOpts field, ensuring custom options are included in all requests. [1] [2]

Transcription Response Handling:
Implemented the client.Unmarshaler interface for the Transcription struct in pkg/schema/transcription.go, allowing it to handle various response formats (JSON, SRT, VTT, plain text) based on content type headers. [1] [2]
Added necessary imports to support content type parsing and response handling.
These changes collectively make the HTTP client more extensible and robust when interacting with transcription and translation endpoints.

Full Changelog: v0.0.36...v0.0.37

Downloading the CLI

Download the CLI for Linux and MacOS here. This does not contain the server. To check the integrity of the binary, and make it executable:

# Download and check against the SHA256 hash on the releases page (MacOS)
wget -O gowhisper -q https://github.com/mutablelogic/go-whisper/releases/download/v0.0.37/gowhisper-darwin-arm64
sha256sum gowhisper
chmod +x gowhisper

# For MacOS, remove quarantine attribute
xattr -d com.apple.quarantine gowhisper

# Run the executable
./gowhisper --help

What's Changed

• Added realtime transcription and translation support by @djthorpe in #89
• Fixed dependencies for Vulkan container

This is a pre-release version, and further changes which can affect the API might occur.

Full Changelog: v0.0.35...v0.0.36

Downloading the CLI

Download the CLI for Linux and MacOS here. This does not contain the server. To check the integrity of the binary, and make it executable:

# Download and check against the SHA256 hash on the releases page
wget -O gowhisper -q https://github.com/mutablelogic/go-whisper/releases/download/v0.0.36/gowhisper-darwin-arm64
sha256sum gowhisper
chmod +x gowhisper

# For MacOS, remove quarantine attribute
xattr -d com.apple.quarantine gowhisper

# Run the executable
./gowhisper --help

• Fixed GGML_NATVE=OFF for vulkan docker builds

Full Changelog: v0.0.34...v0.0.35

Download the CLI for Linux and MacOS here. This does not contain the server. To check the integrity of the binary, and make it executable:

# Download and check against the SHA256 hash on the releases page
wget -O gowhisper -q https://github.com/mutablelogic/go-whisper/releases/download/v0.0.35/gowhisper-darwin-arm64
sha256sum gowhisper
chmod +x gowhisper

# For MacOS, remove quarantine attribute
xattr -d com.apple.quarantine gowhisper

# Run the executable
./gowhisper --help

What's Changed

• Fixing issues with CUDA Docker build by @djthorpe in #88

Full Changelog: v0.0.33...v0.0.34

Download the CLI for Linux and MacOS here. This does not contain the server. To check the integrity of the binary, and make it executable:

# MacOS
wget -q https://github.com/mutablelogic/go-whisper/releases/download/v0.0.34/gowhisper-darwin-arm64
sha256sum gowhisper-darwin-arm64
chmod +x gowhisper-darwin-arm64
xattr -c gowhisper-darwin-arm64
./gowhisper-darwin-arm64 --help

What's Changed

• Prepare for gowhisper release by @djthorpe in #87

Full Changelog: v0.0.32...v0.0.33

Download the CLI for Linux and MacOS here. This does not contain the server. To check the integrity of the binary, and make it executable:

# MacOS
wget -q https://github.com/mutablelogic/go-whisper/releases/download/v0.0.33/gowhisper-darwin-arm64
sha256sum gowhisper-darwin-arm64
chmod +x gowhisper-darwin-arm64
xattr -c gowhisper-darwin-arm64
./gowhisper-darwin-arm64 --help

Full Changelog: v0.0.31...v0.0.32

Download the CLI for Linux and MacOS here. This does not contain the server. To check the integrity of the binary, and make it executable:

# MacOS
wget -q https://github.com/mutablelogic/go-whisper/releases/download/v0.0.32/gowhisper-darwin-arm64
sha256sum gowhisper-darwin-arm64
chmod +x gowhisper-darwin-arm64
xattr -c gowhisper-darwin-arm64
./gowhisper-darwin-arm64 --help

What's Changed

• Re-introduced translation by @djthorpe in #75

Full Changelog: v0.0.24...v0.0.25

Full Changelog: v0.0.23...v0.0.24