Pre-seed · 2026

AI. In your pocket.

A card-sized recorder that turns what you say into your day — written down, summed up, and pointed at what to do next.

The problem

The best thinking happens away from a keyboard.

Everyone already carries a recorder and almost nobody uses it, because a recording is a chore you have postponed rather than a thing you have finished. The idea at the red light, what the client actually agreed to, what the doctor said — the moments worth keeping are exactly the ones where opening an app is impossible.

• Voice memos are write-only. Almost nobody plays them back.
• Phone assistants need a wake word, a screen, and a free hand.
• Meeting tools only work in meetings that were scheduled.

How it works

Capture, then everything after.

1. 01

   Capture

   Press and talk. Works offline; syncs when signal returns.

2. 02

   Understand

   Transcribed, then read for decisions, commitments and next steps.

3. 03

   Act

   Connected to calendar and mail, so a next step becomes an event or a drafted reply.

Why now

The hardware got cheap. The models got good.

A capable always-with-you recorder costs a fraction of what it did three years ago, and the intelligence that makes a recording useful only became reliable in the last eighteen months. The window is the gap between those two facts and the point where a phone maker builds it in.

• Transcription and summarisation are now good enough to trust unattended.
• A dedicated device wins on the one thing a phone cannot do: zero friction, no screen.
• Recurring software revenue on a hardware install base is the model that works.

Sources: OpenAI pricing · Stanford AI Index 2025 · Apple Newsroom, WWDC 2024

The comparable

Plaud proved the category.

Plaud proved people pay for a dedicated recorder and then keep paying for the software — and did it without venture scale: crowdfunding, then profit, with its only institutional money (~$5M) arriving after the first $100M of run rate. Meta has since bought Limitless; Amazon bought Bee. The category is being bought, not questioned — and this round buys a production run and a proven funnel, not a market.

The difference

They summarize. We follow through.

They sell

Meeting notes — a summary of the meeting you sat in, read after the fact, or an always-on microphone that hears everything to catch anything.

We sell

One press, anywhere — and a layer that learns your projects and people, compounds with every capture, and lands the next step in your calendar and your mail.

The edge

The device is the easy part.

Anyone can have a recorder made. Nobody else is building the layer above it — the part that knows which project you are talking about, who you owe an answer to, and what you actually meant to do next. That layer is the company. The card is how it gets into your pocket — and the same layer sells to companies as their own: branded devices, the data resident in their VPC, so leadership sees the work as it happens.

• Context, not transcription

  Transcription is a commodity API. Knowing that "the Henderson quote" is a thing you owe someone on Friday is not.

• It compounds

  Every capture teaches it your projects, your people and your patterns. Month six is materially better than week one, and a competitor starting fresh is back at week one.

• It acts

  A summary is a document. An event on Thursday and a drafted reply are outcomes — and the connections that make that possible are software, not silicon.

The market

Software widens who this is for.

A recorder that only takes meeting notes sells to people who sit in meetings. A layer that adapts to its owner's context sells to anyone whose work happens away from a desk — and each segment reached is the same hardware with a different software posture, not a new product line.

• Founders and managers11.1M US
• Trades and field service12.5M US
• Consultants and agencies1.8M US
• Clinicians1.5M US
• Students and researchers3.3M US
• Field sales1.5M US

Bottom-up, US only: about 32 million seats across these six segments, counted from BLS, AAMC, NCES and NSF data. At $108 a year of membership that is a ~$3.4B software ceiling before any device revenue — and the BLS counts exclude the self-employed, so trades and consultants run higher than shown.

Sources: BLS OEWS · May 2025 · AAMC · 2024 · NCES · Digest 2023 · NSF GSS · fall 2024

The model

$129 once. $9 a month.

Hardware gets it into a pocket. The membership is the business — it runs the transcripts, the daily brief and the connections, and it is what turns a one-time buyer into an install base. The attach rate this banks on has a benchmark: Plaud reports nearly half of its device buyers upgrade to a paid plan.

Device

$129 · $179 Pro

Membership

$9 / month

Unit margin

≈50% — $65 landed on $129

Payback

No CAC yet — this round buys that number

Sources: TechCrunch · Jun 2026 — ~50% attach · Plaud pricing — $159 device, $8.33–$19.99/mo

Where we are

Built, not announced.

The product exists. The manufacturer has produced dimensioned drawings for both models and units have been made. What has not happened yet is going to market — which is exactly what this round is for.

• Hardware

  Two models drawn and produced with the manufacturer

• Brand and site

  Live at hichit.com

• Software

  The app is built — transcription, the daily brief, context-specific accounts, and the calendar and mail connections.

• Demand

  Three signed company LOIs — one in healthcare AI, one in logistics, one in automotive.

Team

Two builders.

The ask

$200,000 for 5%.

Pre-seed · $4M post-money

Ignition, not runway to nowhere. Enough to get a production run out the door and prove the funnel converts.

Go to market

50% — proving the funnel

Production run

20% — the first 100-unit run

Software

10%

Buffer

20%

Half the round goes to go-to-market on purpose: the device is made, the app is built, and the only number this company does not have yet is what a customer costs. Six months of runway to get it.

Instrument: Post-money SAFE, $4M cap