Anker has introduced a new AI-driven voice recorder named the Soundcore Work, which aims to enhance the way users transcribe, summarize, and highlight key moments from conversations or meetings. This latest release competes with devices such as the Bee and Plaud NotePin but distinguishes itself with a compact, coin-sized design that makes it a more portable option.
Slated for launch in September, with an expanding global release anticipated later in the year, the Soundcore Work measures a compact 0.91 inches wide—slightly smaller than a quarter in diameter, yet thicker than a coin. Priced at $99.99, it includes basic functionalities, while an additional subscription service at $15.99 per month will provide users with extra features that have yet to be disclosed.
The Wearable will be offered in black and white color options. Designed for versatility, it can be attached to a lanyard or clipped onto clothing with the included magnetic clip. Despite its featherweight of just 10 grams, the device promises up to eight hours of battery life. Additionally, it comes with a proprietary charger that extends usage time up to 32 hours.
Unlike traditional recorders that continuously capture audio, the Soundcore Work activates its dual microphones only when users tap the device. A double tap allows users to mark significant sections of a conversation, which will be highlighted in the transcriptions later on.
Utilizing the capabilities of GPT-4.1, and eventually transitioning to GPT-5.0, the device can transcribe audio in over 100 languages. Users will have access to both transcriptions and audio recordings through the Soundcore mobile app, enabling them to search or export the content in various formats. When users request transcription or summarization services, the encrypted audio files are sent to the cloud of the AI service provider for processing, with assurances that they will be deleted post-processing. For file sharing, the audio is also temporarily routed through the cloud.
Correction, September 4th: Anker initially indicated that audio recordings would only be uploaded to the cloud upon sharing, but later clarified that they are also briefly sent to the cloud during transcription.
