AI note-taking hardware has spent the past two years trying to find the right shape. We have seen pendants, pins, pocket recorders and phone accessories all promise the same thing: record a conversation, transcribe it and let AI turn it into something useful.

Plaud now thinks the answer may be something people already wear every day.

The company has introduced Plaud One, a pair of AI-focused wireless earbuds built for calls, meetings and in-person conversations. On the surface, the product looks familiar. The more interesting part sits in the charging case: 4G LTE connectivity powered by a built-in eSIM.

That small detail changes the proposition. Plaud is not simply trying to make smarter earbuds. It is trying to give its AI agent a connection of its own.

The case is doing more than charging

Plaud One can record phone calls and online meetings through the earbuds, while the case can independently capture face-to-face conversations. Plaud says the case uses four MEMS microphones with AI beamforming and can pick up voices from up to five metres away.

The earbuds use 12mm drivers, Bluetooth 5.4 and adaptive active noise cancellation of up to 40dB. Plaud lists up to six hours of continuous recording through the earbuds and up to 25 hours through the case. Total listening time with the charging case reaches up to 36 hours.

Those are solid hardware numbers, but they are not what makes Plaud One unusual.

The case has 4G LTE and an embedded eSIM. According to Plaud, that means it can stay connected even when the user’s phone is not nearby. The case also has dedicated record and agent buttons, plus a speaker, allowing users to capture conversations or speak directly to Plaud’s AI.

This is where eSIM stops being a telecom feature and starts becoming product infrastructure.

Instead of asking users to tether the device to a smartphone every time the AI needs the cloud, Plaud can keep the agent connected in the background. For an AI wearable, that independence may prove more important than another improvement in microphone quality.

From transcription to action

Recording is only the first layer.

Plaud says its system supports transcription and summaries in more than 112 languages. The company’s agent can use conversation context to generate notes, answer questions, draft follow-ups and create documents. Plaud is also building connectors for services including Gmail, Google Calendar, Google Drive, Slack and Notion, although some integrations and agent features may arrive through later software updates.

That distinction matters. AI note-taking itself is no longer particularly novel. Software services such as Granola, Fathom and Read AI already turn meetings into summaries, action points and follow-up work. The competition is increasingly moving from “Can AI take notes?” to “What happens after the notes are taken?”

Plaud wants the answer to be: the agent acts.

CEO Nathan Xu described the broader ambition in a briefing with TechCrunch:

“ we need to have a wearable and an always-ready interface, one that people love to wear, is socially acceptable, and private by design.”

That is a more interesting argument than simply selling another recorder. Earbuds are already socially accepted, carried everywhere and sit at the point where voice, audio and personal computing meet.

But people already own earbuds

Plaud’s biggest challenge is obvious.

Most potential buyers already have AirPods, Galaxy Buds, Pixel Buds or another pair of wireless earbuds. Pair those with a capable AI note-taking service and much of the basic recording-and-transcription proposition is already covered.

There is also direct competition. Viaim’s RecDot costs $199.99, records calls and meetings, offers AI summaries, supports transcription in 78 languages and includes 600 transcription minutes per month. It can also start recordings from the earbuds or charging case without relying on a phone for the capture itself.

Plaud therefore needs its cellular independence and agent layer to deliver something meaningfully better, not merely something slightly more convenient.

The economics deserve watching too. Plaud One includes a Starter Plan with 300 transcription minutes and $200 in Plaud credits for eligible Intelligence features. 4G LTE access is included with the Limited Explorer Edition, but Plaud says future cellular-plan details will be announced separately.

That is where the experience could become messy. Embedded connectivity works best when users barely notice it. If owning the device eventually means managing transcription allowances, AI credits and a separate cellular charge, the simplicity starts to erode.

The Explorer Edition reinforces the experimental nature of the launch. Plaud is limiting it to 2,000 units worldwide at $249.99. Its current product page says the first pre-order round has ended and directs interested buyers to a waitlist.

Conclusion

Plaud One is interesting less because it puts AI inside earbuds and more because it gives the AI its own network connection.

Viaim is already proving there is demand for earbuds that record and summarize conversations. Software players such as Fathom, Granola and Read AI are pushing meeting intelligence deeper into workflows. Meanwhile, mainstream earbud makers own the daily listening habit. Plaud is trying to connect those three worlds: familiar hardware, persistent context and an agent that can act without waiting for the phone.

For the eSIM industry, that is the bigger signal. Embedded connectivity is moving beyond smartphones and travel plans into devices where the user may never think about the SIM at all. The network simply keeps the product useful.

That is also the standard Plaud now has to meet. If the eSIM disappears into the experience and the agent genuinely saves users work, Plaud One could point toward a more independent generation of AI wearables. If users end up thinking about connectivity plans, credits and transcription minutes as separate products, the magic disappears quickly.

The winning AI wearable may not be the one with the smartest microphone. It may be the one that makes connectivity, context and action feel like a single thing.