Plaud

Plaud makes a small recorder you clip on or set on a table, and an app that turns whatever it hears into a clean transcript, a summary, and a list of what to do next. That is the whole pitch, and it worked: the company says it crossed a $250 million revenue run rate in roughly three years, and it did it bootstrapped and profitable, with margins its founder likes to compare to Apple's.

That makes Plaud a strange bird in AI hardware, a field littered with well-funded flops like Humane and Rabbit. Founded in 2023 by Nathan Xu and Charles Liu, a Shenzhen wearables factory owner, it leaned on China's supply chain to ship real devices and on a subscription app to keep the money coming after the sale. More than two million people now use it across 170-plus countries, and the company is pushing from solo professionals into teams, healthcare, and enterprise.

Plaud's bet is that the friction of note-taking is a hardware problem, not just a software one. A pin on your lapel or a card on the desk records the room without you opening an app or setting your phone down mid-meeting, and the Plaud app then transcribes in more than 100 languages, writes the summary, and pulls out action items.

The company frames this as "intelligence infrastructure" for what you say, hear, see, and think, which is grander than it needs to be. In practice it is an ambient recorder plus an AI layer called Plaud Intelligence, sold to doctors, lawyers, salespeople, and anyone who lives in back-to-back meetings.

Everyone in a meeting is either present or taking notes, rarely both. Plaud's whole reason to exist is to remove that trade-off: capture the conversation hands-free, and get the transcript, the summary, and the follow-ups without replaying an hour of audio.

The device angle also solves a quieter problem that pure software struggles with. A dedicated mic sitting near the speaker records more cleanly than a phone buried in a bag, and cleaner audio is what makes the AI summary actually usable.

Plaud sells to individuals first: doctors charting patient visits, lawyers logging client calls, salespeople capturing deal conversations, and knowledge workers drowning in meetings. The pitch is personal productivity, one device and one subscription at a time.

The next act is teams and enterprise. HIPAA compliance and healthcare-specific messaging point at regulated verticals, and a "For Teams" plan plus a 0-to-1 B2B product effort signal a real push past the solo buyer into shared workspaces.

The hardware splits by form factor and setting: the Note and Note Pro are cards for meetings and calls, while the NotePin and NotePin S are wearables for hands-free, in-person capture. Prices run from $159 to $189, and every device is a front end for the same Plaud Intelligence app.

The app is where the recurring revenue lives, and it is a genuine product in its own right: transcription in 100-plus languages, AI summaries, an "Ask Plaud" search, thousands of templates, and cloud sync you can turn off if you would rather keep data on-device. Plaud also ships a developer SDK, API, and an MCP integration for building on top of it.

Plaud makes money twice: a one-time hardware sale, then a subscription for the AI transcription and summaries. About half of revenue comes from the recurring app plans, which is what turns a gadget business into a software one with hardware margins on top.

The free tier gives you 1,200 minutes a month; the Pro plan is $17.99/month or $99.99/year at that same cap, and Unlimited is $29.99/month or $239.99/year with no minute limit. A Team plan and per-minute add-on packs round it out. It is a classic freemium-to-subscription funnel bolted onto a hardware DTC brand.

Plaud's rivals split in two. On software there is Otter.ai and Fireflies.ai, meeting assistants that need no extra device; on hardware there was a graveyard of ambitious AI gadgets like Humane and Rabbit, whose failures Plaud partly benefited from by staying narrowly useful instead of trying to replace the phone.

The edge is the device itself: ambient, hands-free capture that a phone app cannot match, paired with a sticky subscription. The catch is the same thing. Asking people to carry and charge another gadget, and to trust one ecosystem with their conversations, is friction that pure-software rivals simply do not have. In China the fight is harder still, with ByteDance and DingTalk shipping their own recorders.

The praise is consistent: on clean audio the transcripts are accurate and the AI summaries genuinely save time, and reviewers like the discreet, hands-free capture far more than propping up a phone. For meeting-heavy professionals, the after-the-fact clarity is the whole point.

The complaints cluster around reliability and lock-in. Some users hit failed uploads, sync delays, or connection issues, which sting because the product depends on cloud transcription, and a few report hardware wearing out after months. The other recurring gripe is philosophical: it is one more device to carry, charge, and tie yourself to.

Plaud is the rare AI company that barely raised money. It started with a roughly $1 million Kickstarter in 2023, took a small check of around $5 million from Carbide Ventures, and otherwise funded itself, so the founders still own most of it. Revenue did the heavy lifting: about $56 million in 2024 at a roughly 20% margin, then a near-tripling to a reported $250 million run rate in 2025.

The valuation is murkier. Reports pegged a mid-2025 Tencent investment at a $1 billion valuation, later chatter put it near $2 billion, and both Plaud and Tencent have called the reports untrue. Treat the number as unconfirmed; the profitability is the part that is actually documented.

Plaud has the enviable problem of a working business: profitable, growing fast, and expanding from individual users toward teams, healthcare, and enterprise, with a stated ambition of reaching tens of millions of professionals. The subscription attach rate is what makes any of it durable.

The risks are the ones a hardware-plus-cloud company always carries. Reliability complaints, the friction of yet another device, deep-pocketed rivals now entering the category, and a security and business-systems program still being built out are all real. The next chapter is whether the software and the enterprise motion mature fast enough to match how quickly the hardware sold.

Plaud is distributed by design, with engineering, product, and go-to-market teams across San Francisco, Shenzhen, Beijing, Tokyo, Singapore, and Europe. Founder Nathan Xu based hardware in Shenzhen deliberately, to sit next to the supply chain, and pairs it with a US commercial front. The stated values run toward first-principles thinking, using AI to the fullest, and "human first," and roles are hybrid at three days a week in office.

Culture here reads as high-velocity and AI-native. Engineering and analytics postings openly expect people to build with Cursor, Claude, GPT, and Gemini as the default, and the recurring words are ownership, craftsmanship, and speed. Many roles are framed as founding-team seats on brand-new initiatives, which is the flip side of a young company still standing up its systems and its security program from scratch.

Plaud discloses pay on many of its US roles, all in dollars and annualized, and the range is genuinely wide. Engineering lands roughly $140K to $285K on most postings, with design and marketing bands sitting a notch lower, while the standout is the Speech LLM lab, where a couple of research roles stretch to $200K to $540K base.

Compensation is a package, not just base. Roles carry an Employee Stock Ownership Plan or equity plus a performance bonus, so the equity upside in a bootstrapped, profitable company is a real part of the story, not a footnote.

For a product built to record private conversations, trust is the whole game, and Plaud stacks the certifications to prove it: SOC 2, HIPAA, GDPR, ISO 27001, ISO 27701, and the EN 18031 hardware standard. The company says data is encrypted in transit and at rest, is never sold or used to train its models, and can be kept on-device rather than in the cloud if you choose.

The registered entity is Plaud Inc., at 8 The Green, Ste A, Dover, Delaware, a Delaware-incorporated company operating out of San Francisco. Security is clearly a live buildout rather than a finished one: recent job postings describe standing up the SIEM, CSPM, incident response, and SOC 2 Type II evidence largely from zero.

Plaud's coverage tracks one arc: an under-the-radar hardware maker that turned profitable while flashier AI startups stalled. Forbes framed it as one of the few AI companies actually making money, and Asian tech press has chronicled both its supply-chain playbook and the on-again, off-again Tencent rumors.

The links below are the source trail for the claims on this page, from founder interviews to the funding reports the company disputes.