This paper is about one specific kind of call: two people, high stakes, detail that matters. A founder pitching an investor. A reviewer judging finish and texture. An interviewer reading a face. In every one of these, the call is not a meeting — it is a showing. And the standard tools were built for meetings.
I 4kChat because the pitch has to be shown
Some businesses survive a blurry call. A status update does. A standup does. But if your work has to be seen — the grain of a material, the fit of a prototype, the expression on the other side of the table when you name a number — then the call itself is part of the product. You are not describing the thing. You are putting it in front of someone.
That is the through-line of this paper and of the whole High-end 1:1 calls cluster: a high-end 1:1 is a presence problem, not a meeting problem. The question is never "did the call connect?" It is "did the other person actually see what I needed them to see?"
What "you're in the room" means on a real display
Presence is not a metaphor. On a real display — a large monitor or a TV at viewing distance — a high-resolution feed of a person is roughly life-size. Eyes are where eyes are. You can see someone lean in. You can see the moment a reaction happens, not a smeared approximation of it a frame later.
Shrink that same person to a two-inch tile in a grid and the information collapses. Micro-expressions disappear first. Then texture: the difference between brushed and polished, between matte and gloss, between "handmade" and "printed." Then posture — the small shifts that tell you whether to keep going or stop. What remains is a talking avatar. Fine for a sync. Fatal for a showing.
"You're in the room" is the standard we hold a call to: could the other person judge the thing — the product, the face, the room — as if they were there? If the answer is no, the resolution, the frame rate, or the display path failed you, and no amount of meeting features compensate.
Where Zoom/Teams compress presence out of the call
Mainstream meeting tools are engineered for scale: dozens of participants, mixed networks, screen shares, recordings, transcripts. To make that work reliably, every video stream travels to a central server, gets decoded, re-encoded, and forwarded. That pipeline is the right call for a forty-person all-hands. It is also where presence dies.
Each encode step spends a bitrate budget, and the budget is set for the worst case, not your case. The encoder prioritizes motion smoothness and legibility of slides — not the finish on a product or the texture of skin and fabric. Fine detail is the first thing a rate-controlled encoder throws away, because it is the most expensive thing to keep. Add a tile layout that renders each person at a few hundred pixels wide, and even a perfect upstream feed is displayed as a postage stamp.
None of this is a criticism of those tools at their job. It is a statement about fit: a server-routed, tile-gridded, scale-first architecture optimizes for many people hearing each other. A high-end 1:1 needs the opposite — two people seeing each other, at the highest fidelity the hardware and network allow. The full side-by-side lives on the comparison page.
Peer-to-peer: why no media server helps the picture (and what it costs)
4K Chat connects the two browsers directly. There is no media server in the path: nothing decodes your video, re-encodes it, or decides what quality you "should" have. The camera's feed is encoded once, on your device, and decoded once, on theirs. Every bit of detail the encoder kept is the detail the other side receives.
That removes the two biggest taxes on a 1:1 picture: the server re-encode, and the shared bitrate policy of a big meeting platform. It also means your audio and video never sit on someone else's infrastructure — the privacy story is the architecture, not a policy page. See how it works for the full picture.
The cost of this design is honesty about connectivity. Direct peer-to-peer needs a path between the two networks, and some corporate or carrier-grade NATs block that. The Free tier uses STUN only — there is no TURN relay — so on a sufficiently strict network, a Free call may fail to connect rather than fall back to a relay. Paid plans offer a bring-your-own-relay option that you operate yourself; we still do not put a media server in your call. We would rather tell you this up front than quietly route your "peer-to-peer" call through our servers.
Receiving life-size: the cast/mirror path
The sending side is only half of presence. If the other person watches you on a laptop balanced on a stack of books, you are a postage stamp again — just a sharper one. The fix is a big screen, and the honest way to get there today is casting or mirroring: the call runs in the browser on a phone or PC, and that device casts or mirrors to the TV.
There is no 4K Chat app in any Smart TV store, and we do not claim one. The browser is the path. In practice this is not a compromise: the TV is a display, the browser does the work, and the result on the wall is a life-size person. The step-by-step is in the big-screen guide and the pitch-on-a-4K-TV tip.
Setup for founders, reviewers, interviewers: the 60-second start
The workflow is deliberately unremarkable, because the call should not be the memorable part. Open Start a Call, allow the camera, and you have a room. Send the link or the code. The other person opens Join — no account required, camera optional — and lands in the waiting room, which is on by default. You see who is asking to come in, you admit them on your cue, and the call begins.
That waiting room matters more than it looks. It means you can be framed, lit, and composed before the investor sees anything. It means a reviewer never accidentally drops into the previous call. It means the first thing the other side sees is the version of the room you chose. The details are in the waiting-room guide.
Gear that actually moves the needle
In order of impact, for a high-end 1:1:
Light first. Soft, even light in front of you does more than any camera upgrade. Face a window or a soft lamp; never sit with a bright window behind you — the camera exposes for the background and your face becomes a silhouette. See the lighting tip.
Camera second. A recent laptop camera is usable; a decent external webcam or a capture-connected camera is better. What matters is a clean, well-lit sensor image — resolution claims on the box matter less than low-light behavior.
Audio third. People forgive soft video; they do not forgive bad audio. A headset or a simple USB microphone beats the laptop mic, mostly by getting the mic closer to your mouth than to the keyboard.
Screen last, on the receiving side. The biggest single upgrade to received presence is the display the other person watches you on. The full list is in the gear checklist.
Limits that keep trust
Seats are caps, not suggestions: Free up to 4 participants, Pro up to 12, Cinema up to 32. For a 1:1, every tier has the seats you need; the tiers differ in observers, cameras, and recording, not in whether two people can talk. Current numbers live on Pricing.
Resolution follows your hardware and network. We do not guarantee 4K or 8K on every call, and any product that does is marketing, not engineering. What the architecture guarantees is that nothing in the middle throws detail away: whatever your camera, encoder, uplink, and the receiver's screen can sustain is what the call delivers. If your network can only carry 1080p that day, you get an honest 1080p — not a "4K" label on a compressed stream.
When Pro beats Free for a 1:1
For a pure two-person call, Free is genuinely enough — that is the point of the tier. Pro earns its place the moment the 1:1 grows an audience or an angle: a couple of observers sitting in on the pitch, a second camera on the product table while the first stays on your face, or local recording of the session. If your calls are strictly two people, one camera, no recording — stay on Free and spend the money on a light.
Try it: measure received detail, not marketing claims
Everything above is testable in one call, and you should test it rather than believe it. Open Start a Call, have the other side join, and put something with fine detail in frame — a product, a printed page, a fabric. Cast or mirror the receiving end to the biggest screen you have. Then judge: can they read the texture? Can you read their face? That measurement, on your hardware and your network, is the only benchmark that matters for your calls.
If it falls short, the fix is usually in this paper: light, camera position, the receiving screen, the network — or, on a strict network, the reality that Free has no relay. The postage-stamp mistakes page collects the common failure modes, and Support has the how-tos. Native apps are in development; today, the browser is the whole product, and it is enough to find out whether presence is worth more to you than meeting features.