Pocket makes an AI meeting recorder that snaps onto the back of your phone and turns real-world conversations into transcripts, summaries, and tasks. The startup has now raised $11 million from Accel, Y Combinator, and ElevenLabs CEO and co-founder Mati Staniszewski. Offline meetings are still messy to capture cleanly, and most people are still patching together smartphones and note-taking apps that work fine, but not especially elegantly.
Founded in 2024 by Akshay Narisetti and Gabriel Dymowski, Pocket is trying to prove there’s room for one more hardware startup in a category littered with half-baked AI gadgets. Its bet is simple: make the thing cheaper and nicer to carry. Then make it useful enough that people keep using it.
What is Pocket AI recorder and how does it work?
Pocket’s device is basically a slim aluminum puck that lives on the back of your phone. You press a side button to start recording. Flip a physical slider to switch between phone-call mode and in-person conversation mode. Then sync the audio into the app through Bluetooth, Wi‑Fi, or USB‑C. For calls, Pocket uses a contact mic while attached to the phone, so it doesn’t rely on speaker mode. For heavier use, its Wi‑Fi transfer mode can move files up to 20x faster than Bluetooth.
Once a recording lands in the app, Pocket runs it through a clear pipeline: upload, analyze, extract, then write. The software identifies speakers and context. It pulls out themes and action items, then formats the result into a structured note instead of a giant blob of transcript text. Users can choose different summary lenses like Meeting, Interview, or Journaling. A sales call and a lecture shouldn’t be summarized the same way.
The app goes further than plain transcription. Ask Pocket lets users query their recordings in chat form and filter by date or conversation. It also lets them revisit earlier sessions. Pocket also auto-generates mind maps from summarized conversations and adds tasks to a calendar with one tap. Its API exposes semantic search so teams can find relevant material across recordings, transcripts, and summaries.
For enterprise customers, the company offers custom workflow management and webhook support. It also integrates with Google Calendar, OneDrive, Google Drive, Obsidian, Claude, and Cursor. Its MCP server lets compatible AI tools pull transcripts, summaries, date-ranged conversation searches, and action items directly into outside workflows.
Who founded Pocket and why did they build it?
The founding story
Pocket was founded in 2024 by Akshay Narisetti and Gabriel Dymowski, and the company is now based in San Francisco. Narisetti’s pitch is that most AI note-taking products were built for Zoom-era conversations, while a lot of valuable context still lives in hallways, client meetings, classrooms, job sites, and phone calls. Pocket’s reason for existing comes down to that gap. The company is active, part of Y Combinator, and lists a team of 15.
Narisetti put it plainly in the source interview: “We thought every meeting notetaker was built for online conversations, but nothing was geared towards real-life talk. AI really needs a lot of context to work better for us, and a lot of that context exists offline.”
Founder market fit
There’s real category overlap here. Narisetti was a founding member of Omi, which also works on note-taking, so he wasn’t walking into the category cold. Dymowski, meanwhile, previously built a blockchain-based document management startup. He came in with experience around structured records, storage, and how people handle business information once it’s been captured.
That doesn’t guarantee anything.
But it does make Pocket’s founding team more credible than the usual “we discovered meetings are annoying” startup duo.
Traction, pricing, and early proof
Pocket has sold more than 130,000 units since launching last year, which is a meaningful signal in a hardware category where a lot of products get attention and then quietly disappear. The device sells for $129 and is pitched with unlimited recordings, transcriptions, and to-do items without a required subscription for the core experience.
The monetization layer sits above that. Pocket also sells a $200-a-year plan that unlocks unlimited AI summaries and unlimited chat with Ask Pocket. It also includes daily highlights, file attachments, advanced templates, priority processing, and support for recordings up to 4 hours long. That split is smart. The hardware gets users in, and the paid software tier tries to turn occasional recording into a habit.
The funding round and what it says
The new round totals $11 million, with Accel, Y Combinator, and Mati Staniszewski backing the company. For a startup that only launched last year, that’s a strong vote that investors think the opportunity isn’t just selling a nicer recorder. It’s owning the layer that turns ambient conversation into searchable work.
How does Pocket AI recorder compare to Plaud, Otter, and phone apps?
Pocket doesn’t have the market to itself. On the hardware side, it’s up against Plaud, Mobvoi, Anker, Viaim, and Vibe. On the software side, it runs into Granola, Zoom, Fireflies, Otter, and Read AI. The most obvious incumbent is still the phone already sitting in your pocket with a recording app and a notes app.
That last point matters.
A dedicated recorder has to earn its keep because the fallback option is cheap and good enough. Pocket’s differentiation is the physical design and the $129 price. It also has the phone-backed form factor, offline capture, and a no-subscription core product that doesn’t immediately punish casual users. It also understands that hardware alone won’t win. That’s why it keeps shipping API hooks, enterprise integrations, and MCP support instead of stopping at “cool gadget.” Plaud is already building enterprise capacity and desktop support for digital meetings, and the source article says it’s on track for $100 million in annual revenue via software sales. Pocket is chasing the same truth: the durable business is probably software, even if hardware gets the customer in the door.
Why are investors backing Pocket AI recorder now?
Accel’s thesis is less about transcription and more about memory. Cecilia Wang, a partner at the firm, argued that users stay present in meetings instead of splitting attention between listening and taking notes, and that the accumulated record of conversations becomes increasingly valuable over time.
That’s the bull case.
Pocket isn’t just selling a recorder. It’s trying to become a personal archive of conversations that can later be searched, summarized, connected to calendars, or piped into other tools. That’s a much bigger ambition than basic note capture, and it’s why this round matters. Money like this gives Pocket room to keep pushing its software faster, especially the enterprise features, while still using the device as the wedge.
The investor mix is telling too. Accel brings the classic early-stage software lens. Y Combinator gives the company startup-distribution muscle. Staniszewski’s backing fits a broader belief that voice interfaces and speech-native workflows are getting a lot more important, not less.
How big is the AI meeting assistant market?
Pretty big — and growing fast. The global AI meeting assistant market was valued at $3.5 billion in 2025 and is projected to reach $21.5 billion by 2033, which implies a 25.8% compound annual growth rate. North America held the largest regional share at 33% in 2025, and software accounted for 70% of the market.
Those numbers line up with what’s happening. Companies want real-time transcription and multilingual support. They also want workflow automation and better integration with collaboration tools. They care a lot more now about secure and compliant handling of recorded conversations. That trend helps Pocket, but it also raises the bar. A cute device won’t be enough.
Pocket AI recorder has a real shot — if the software keeps up
Pocket’s story is easy to understand: a tidy hardware product, a lower entry price, early sales traction, and investors willing to fund the next layer. But the harder part starts now. The AI meeting recorder market won’t be won by whoever ships the prettiest puck. It’ll be won by the company that makes captured conversation instantly useful at work — and keeps doing that faster than software-only rivals can copy the habit.
Read how AllHome raised ₹200 Cr in a Series B round led by Bessemer Venture Partners to expand its tech-led home improvement platform, adding experience centres, manufacturing capacity, and software for India's fragmented interiors market.
FAQ
- What funding did Pocket raise? Pocket raised $11 million from Accel, Y Combinator, and ElevenLabs CEO and co-founder Mati Staniszewski. It’s an early but meaningful round for a company that launched only last year and has already moved beyond a simple hardware pitch into software and enterprise tooling.
- How does Pocket’s recorder actually work? It works by attaching to the back of a phone, capturing calls or in-person conversations, and syncing audio into Pocket’s app for transcription and AI processing. The app then turns recordings into summaries, action items, mind maps, and searchable notes, with extra automation features like calendar actions and API access.
- Who founded Pocket? Pocket was founded in 2024 by Akshay Narisetti and Gabriel Dymowski. Narisetti previously helped build Omi, while Dymowski had already started a blockchain-based document management company, giving the team a direct background in note-taking and recorded information.
- What market is Pocket competing in? Pocket is competing in the AI meeting assistant and AI voice recorder category. That market was worth $3.5 billion in 2025 and is forecast to hit $21.5 billion by 2033, which explains why both software players and hardware startups are rushing in.
