Feishu recently introduced an unusual piece of AI hardware developed in partnership with Anker Innovations:a wearable AI recorder small enough to resemble a bean.。
Anker led the hardware development,
while Feishu provided the underlying software and AI capabilities.The device connects directly to the Feishu ecosystem, allowing recorded material to flow into the same environment as the team's other work.。
That integration is the most important part of the product, as we will explore below.
GeekOnUp is an intensive Feishu user.
We use it for communication, documentation, project collaboration, and organizational knowledge.
To borrow a phrase that circulated recently, we could describe the condition as
a serious dependence on Feishu.
We have even created an internal practice around it:
colleagues actively discover useful Feishu capabilities and share them across the company.
We organized an internal initiative
called the Feishu Productivity Star。
to encourage that behavior. So when Feishu released a wearable recorder that automatically generates AI meeting notes,
we bought one immediately and used it for several days.
01
What is the device?
It is genuinely tiny.
The unit weighs only10gabout as much as a broad bean and is close to a coin in size.
It comes with a charging case similar in shape to an AirPods case.
A magnetic button design lets users attach it to a collar, cuff, or other convenient position.
Its unobtrusive presence suggests a different standard for intelligent recording devices.
02
How simple is it to operate?
Simple enough to require almost no learning.
Press once to begin recording and press again to stop.
The audio then enters Feishu automatically and becomes an AI-generated summary.
There is no need to take out a phone or open an app.
The conversation that just happened is preserved in full.
Recording becomes an ambient capability that travels with you.
03
What it feels like after several days
Strength 1: a physical button changes behavior
It is easy to underestimate the value ofpressing once to begin recording.That small interaction matters.
Recording on a phone is possible, but usually requires you to
unlock the device, find an app, open it, locate the recording function, and start.
Starting meeting notes in Feishu is also straightforward, but still requires you to
open Feishu, create a meeting, and begin recording.
With this recorder, the process is one press.
The difference becomes especially clearin unexpected situations.There is no preparation time.
As a result, you are more likely to capture the conversation.
The hardware itself is almost unnoticeable.。
The magnetic attachment adds little perceptible weight, while the charging case extends battery life beyond 30 hours. Opportunistic charging largely removes battery anxiety.
Two built-in microphones capture audio from a range of roughly five meters.
The sound field cannot match professional recording equipment,
but it is more than adequate as a voice-to-text capture tool.
Accessories included with the recorder
Strength 2: the recording does not disappear afterward
Where does the material go when a recording ends?
Many teams have a black hole for captured material.Audio sits in a phone, folder, or forgotten corner of cloud storage. A few days later, no one remembers what it contains, much less reviews or reuses it.
Feishu's advantage is that it already contains the organization's working context.
The recording can be associated with people, time, and collaboration relationships, then searched, cited, and revisited.
It stops being an isolated file and enters the operational system directly.
That is why we describe the product asa system-level integration:it makes information easier to incorporate into the organization's knowledge flow.
An AI summary generated from the wearable recorder
There are some limitations.
- The summary is generated automatically after recording. Without background information, the AI must infer context from the conversation alone, so some details can be inaccurate.
- AI summary capacity is also limited under the subscription model, which matters for intensive use.
The current plan works as follows:
- the initial purchase includes six months of membership, 1,200 transcription minutes, and 25 generated summaries.
- After the introductory period, the subscription costs RMB 69 per month and includes 3,000 transcription minutes and 50 summaries.
- When transcription and summary quotas are exhausted, basic recording and data synchronization continue to work.
Recordings under ten minutes also do not consume a summary credit.
The experience could be more flexibleif users could decide manually whether each recording should generate an AI summaryinstead of consuming the quota automatically.
04
Why does a dedicated recorder matter?
Feishu already provides AI meeting summaries, so many users will reasonably ask what this separate device adds.
We had the same question at first.
Before using the recorder, our workflow looked like this:
open Feishu, start a meeting, begin recording, stop, and wait for the summary.
That is already a capable workflow. Feishu performs reliably in both transcription accuracy and structured summarization.
Its practical limitation is that it behaves primarily asa capability for formal meetings.。
Many valuable pieces of information do not emerge in formal meetings.
They may be two sentences exchanged in a hallway, a critical decision discussed during an in-person customer visit, an idea raised casually by a colleague, or a thought you articulate while traveling.
These moments often contain significant value,but the effort required to begin recording is high enough
that we decide to remember them instead.
Human memory is rarely the dependable option we imagine it to be.
05
The wearable recorder fills precisely that gap.
Within our team,capturing conversations has always been frequent but inefficient.
Anyone working in a communication-intensive team will recognize the pattern.
Meetings, customer discovery, internal solution reviews, and quick alignment discussions all carry a high density of information.
Before May last year, we used the following process:
record on a phone or Apple Watch, import the audio into a third-party transcription tool, then use AI to structure the notes.
We later switched toFeishu's AI meeting summaries.
Initially, only our project managers received access.
By January this year, every colleague had a seat.
We also introduced asmall internal rule:when a conversation is likely to matter, start AI meeting notes whenever practical.
Over time, that created a habit that was both useful and slightly amusing.
The following scene has happened more than once:
two colleagues are talking at their desks, realize that the next part is important, and one suddenly says,
'Wait a moment. Do not continue until I start Feishu Notes.'
The interaction reveals an important limitation.
Starting a recording still depends on the moment when someone realizes that the conversation should be captured.
The wearable device moves that threshold one step earlier.
Traditional collaboration depends heavily on manual information processing, separates different media types, and often lacks a coherent approach to knowledge retention.
AI is particularly effective at processing large volumes of fragmented, unstructured material.
A device like this is fundamentally a low-friction, reliable entry point for capturing context.
The easier and less noticeable capture becomes, the more authentic material an organization can make available to AI.
The more complete the context, the greater AI's potential to support collaboration and decision-making.
Previously, information had to enter the Feishu interface before it could enter Feishu's knowledge system.
Now a single button can give a real-world conversation a path into that system.
When recording no longer needs to beconsciously remembered,，
the most immediate and easily overlooked information is more likely to be preserved.
06
Conclusion
Viewed as a standalone product, it is a lighter and more intelligent recording button.
Viewed within Feishu's broader platform, it is attempting something more consequential:
using a hardware touchpoint to capture a more complete form oforganizational memoryso AI can create value from authentic working situations.
Recording shifts from a discrete action into a habit.
Whenrecordingno longer feels like a deliberate, interruptive task,
and capturing a conversation requires only one press on a collar-mounted button,
the most complete information can be preserved without the act of recording disrupting it.
At that point,information is no longer an isolated archive.，
It enters a continuously movingstream of organizational knowledge.。
AI can draw from that stream, connect fragments into a map, and distill conversations into useful insight.Past details can be recalled, connected, and reused when the organization needs them.
Feishu's knowledge Q&A already allows users to ask questions against previously stored documents.
AI can therefore evolve from a tool that executes commands into an organizational second brain that carries collective memory and supports collective thinking.
Frictionless capture gives that system raw material; the movement of knowledge makes it useful.
The larger idea is that
knowledge begins to flow when the act of recording recedes into the background.
When knowledge continues to flow, AI can begin to function as an organization's second brain.
This wearable AI recorder will not suit everyone.
For occasional audio capture, a phone or watch is already sufficient.
For teams that communicate, collaborate, and review decisions continuously, however, it represents a meaningful direction:
AI is moving beyond a feature and becoming a foundational work habit grounded in real operating context.
That shift is worth watching.
