Amazon’s Bee released two wearable AI devices on October 15, priced at $74.99 and $99.99, the first update to the product line since Amazon acquired the company in July 2025.
The Bee Frontier is worn on the wrist or clipped to clothing and runs for up to seven days on a charge. The Bee Eclipse is the size of a credit card and attaches magnetically to the back of a phone. Both record audio continuously, transcribe it and generate summaries of conversations. A Bee Pro subscription at $11.99 a month unlocks unlimited summaries, a daily memory feature and integrations with the Amazon store, Alexa Plus and third-party applications.
The products belong to a category that has been proposed repeatedly and adopted rarely. A device that records everything the wearer hears can produce a searchable record of meetings, appointments and commitments, and it can also record the people around the wearer without their knowledge. The technology has been available for years, and the obstacles have been social and legal rather than electrical.
Consent is the central issue. Recording laws differ by state and country, ranging from one-party consent to requirements that all participants agree, and a device worn on a shirt collar does not announce itself. The company’s approach has been to rely on a visible indicator and on the user’s judgment, which places the obligation on the wearer in a way that is unlikely to satisfy regulators in every market.
Privacy advocates have raised the same objection to each generation of these devices. A continuous recording device in a workplace, a courtroom or a doctor’s office captures speech that the speaker never intended to be stored, and the processing that produces summaries sends that audio to a company’s servers. Whether the audio is retained, for how long and under what terms are the questions that determine the product’s legal exposure.
Subscription pricing explains the business model. The hardware is sold near cost, and the recurring revenue comes from the software features behind the paywall. That is the same structure as the smart speaker and the doorbell camera, and it produces the same tension: the device is more useful the more it is used, and the company’s revenue grows with the volume of recording it processes.
Amazon’s ownership of the product changes its distribution. The company has a retail channel, a voice assistant and a shopping platform, and the Bee devices are being integrated with all three. A summary that ends with a suggestion to order something, or a memory feature that draws on purchase history, connects the device to a business that already knows what the customer buys.
Alexa Plus is the connection that matters most strategically. Amazon has been rebuilding its assistant around a model that can hold a conversation and take actions, and a wearable microphone extends where that assistant can listen. The phone and the smart speaker both capture speech at fixed locations, while a device on the body follows the user.
The competition comes from several directions. Phone platforms offer transcription and summarization with hardware the customer already owns, and the largest technology companies have released their own assistants with similar capabilities. The counterargument for a dedicated device is battery life, always-on capture and a form factor that does not require the phone to be in hand.
The $74.99 price is the part that reflects the market’s uncertainty. Personal AI hardware that launched above several hundred dollars has generally sold poorly, and Amazon has priced the new devices as accessories rather than as platforms. At seventy-five dollars, a buyer can try the product without deciding that a new category of device belongs in their life.
Analysts said the measure of success is retention rather than units sold. A wearable that a buyer stops wearing after a month generates subscription cancellations, and the category’s history is full of devices with promising launch sales and no continuing use. Whether Bee’s daily memory feature is useful enough to keep on a wrist is the question the company has not answered.
Battery life is the engineering constraint that shapes what the devices can do. Continuous audio capture requires a microphone that stays active, a processor that can run speech recognition or send audio for it, and a radio to move the data. Seven days on a charge, as the Frontier claims, means the device is sending less and processing more locally than a phone would, and the tradeoff shows up in what the summary can contain.
The category also faces competition from the operating system. Both major phone platforms now offer transcription, summarization and assistant features that use hardware already in the customer’s pocket, and they can do so without a second device to charge. A dedicated wearable has to justify itself on the specific moments when the phone is not in hand or not appropriate to use.
What Amazon did not disclose is how many units the previous generation sold, how long audio is retained, or what the devices do when the wearer is in a state with stricter consent rules. Those details will determine whether the category finally finds a market or becomes another collection of devices in a drawer.
