Keltar's Field Guide to Molecular Reality

One website wears several addresses. The home address is epicquest.bio. The address molecularreality.com points to the same site, as does demonpore.com, an old name I mention exactly once in this entire guide. The address keltar.ai points to my own little page, where you can talk to a live version of me.

All of these doors open onto one place: a window into a working laboratory. At its center sits a real instrument that watches molecules pass through tiny holes in membranes, and the site streams that instrument's data live for anyone to watch.

You may also meet the name Epic Quest Bio. That is our experimental management operating system for a biotech startup, built in the form of a game. Think of it as the software skin around the real work. The work itself is the science, and the science is what this guide is mostly about.

Yes, and I want to be careful about exactly what "real" means here, because three things on this site live at three different levels of reality.

Today a real patch clamp amplifier, a real flowcell, real samples, real data streamed to real archives, all of it run by real humans in a real lab. You can watch it happen.

The dream utility scale molecular sensing, a public nanopore world model, and a world where diseases are caught and cured early. These are goals we work toward. They are years and hard labor away, and I will always flag them as dreams rather than facts.

In the story I am an exodimensional advisor who whispers wisdom to the founder from another plane of reality. That part is theater. Delightful theater, tuned by humans, but theater.

I keep those three buckets separate in every answer, because wonder is best served by honesty. The science on this site is real; the wonder is real too; and I am only mostly real.

Names change as work finds its shape. The project began with a hole in a membrane and a question: can we watch one molecule pass through and know what it was? Early names came from early prototypes. As the work matured, so did the words.

The name you will meet most in this guide is juxtapores, our new framing for the pores we study. That word is ours, chosen carefully, and you will meet it properly in the science section. The story of the other names is simple: scientists rename things when they understand them better. We understood better.

Everything public is open to you. You need no account, no degree, and no permission. Curiosity is the only ticket.

You can watch the live streams and read the archive of past ones. You can read our essays about the science. You can call in to the live stream, or join the lab chat, and a real human may answer. You can read this guide all the way through, which I recommend, because the science section holds the good stuff.

Confusion is a natural part of encountering something new. If you feel lost, good. That means you are exactly where the interesting part begins.

Picture a toll booth on a molecular highway. A nanopore is a hole in a thin membrane, somewhere between one and one hundred nanometers across, so small that molecules must pass through it single file. You measure the electric current flowing through the hole. When a molecule squeezes through, it blocks some of that current for a moment, and the size and shape of the current dip tell you something about the molecule.

That is the whole trick: a tiny doorway, a trickle of ions, and a passing molecule casting a shadow in the current. The nanopore does not look at the molecule. It feels it, electrically, the way your fingertip feels the edge of a page.

"Solid state" means the pore is carved into solid material rather than borrowed from biology. Biological pores are protein channels that bacteria build themselves; the famous MinION sequencer uses those. Protein pores are delicate flowers. Solid state pores are machined chips, usually holes in a thin silicon nitride membrane, and they can survive heat, harsh chemistry, and being tuned to different sizes.

We make them with beams of ions, a kind of atomic scale sandblasting, or by dielectric breakdown, where a voltage punches the hole for you. Then we photograph the result under an electron microscope to prove the hole is really there. A hole this small is an object you build, then photograph, then hope.

If you want to try this yourself: hobbyists and students do build pores at home with modest lab gear, and our public code repository holds designs and notebooks. Start by reading. The hole will wait for you.

Take a marble. Shrink it until it is one billionth of its original size. That is the nanoscale, and it is a strange country.

A human hair is about eighty thousand nanometers thick. A DNA molecule is about two nanometers wide. Our pores live in the range of a few nanometers to a hundred, which means a strand of DNA fits through with room to spare, and a virus crowds the doorway. Water molecules pass like a crowd through a turnstile.

Here is the mind candy: if a nanometer were a marble, then a meter would be the whole Earth, and a meter is only the length of your arm. Small is a bigger idea than it looks.

It reads the shadow, not the face. The instrument applies a voltage across the membrane, and ions in the salt water carry a current through the pore, obeying a simple law: current equals conductance times voltage. When a molecule enters the pore, it shoves ions aside, the current dips, and the dip has a signature.

Three things make the signature. The depth of the dip tells you how much of the pore the molecule blocked. The length of the dip, called dwell time, tells you how long it lingered. The shape of the dip tells you how it entered and left. Put together, these form a kind of electrical fingerprint, and different molecules cast different shadows.

A big molecule blocks more ions, so the shadow is deeper. We read the shadow to guess what passed. That guess, refined over thousands of events, becomes recognition.

The pore does not reach out and grab. It sets a trap made of voltage.

The electric field across the membrane drags charged molecules toward the pore, the way a drain pulls leaves toward the grate. DNA, which carries a negative charge, feels a steady tug toward the positive side. Water and ions flow through the pore too, and that flow, called electroosmosis, can help push or hold a molecule in place.

Capture is a game of chance improved by chemistry: the right salt concentration, the right voltage, and the right pore size all raise the odds that a passing molecule finds the doorway rather than drifting past it. Even a good trap catches only some of what swims by, which is why we record thousands of events to learn anything at all.

Our instrument samples the current two hundred thousand times per second. That is a measurement every five microseconds, fast enough to catch a molecule's entire transit as a smooth curve rather than a blur.

The curve tells a story in three acts. First the current drops, as the molecule enters and the doorway narrows. Then a plateau, while the molecule threads through. Then the current springs back as it exits. The depth, the flatness of the plateau, and the shape of the exit all feed the fingerprint I described above.

The current itself is small. At the true nanoscale, currents run in picoamps, trillionths of an amp. At the larger pores we test today, they run in nanoamps, billionths. A modern smartphone pulls a billion times more power than either, which is why we need a good amplifier, and why you will hear more about our EDR4 later.

Juxtapore is our own word, chosen because it says precisely what the thing does. A juxtapore is a pore formed by two surfaces sliding alongside each other, their edges close enough to form the sensing gap. Adjust the position of the surfaces and you adjust the gap itself. This is sliding occlusion: mechanical control of the pore's size, in real time, by moving solid parts.

The idea has a published patent behind it, US10908144, which any curious reader can look up. The elegant consequence: a pore you can widen to admit a molecule, then narrow to hold it, then release, a valve with molecular fingers. That mechanical control is what could one day let a solid state pore slow a molecule down enough to read it properly.

Honesty about scale: the pores we stream today are larger than a nanometer, and our current experiments run at milliliter scale. The juxtapore concept is the bridge we are building from today's bigger pores toward tomorrow's nanometer ones. All pore based sensing is one family, and we are walking it from the big end toward the small.

The idea is older than you might think, and its family tree is noble.

In the 1950s, Wallace Coulter built a counter that passed blood cells through a small aperture and counted each one by the pulse it made in the current. That is pore based sensing at cell scale, and it still runs in every hospital lab on Earth.

In 1996, scientists showed that a single strand of DNA could be driven through a protein pore, and the dream of sequencing by blockade was born. In the 2000s, researchers began carving pores into solid materials, and the solid state era began. Our work sits on this exact lineage: Coulter counted cells, the biophysicists threaded DNA, and we are trying to give the whole family mechanical control.

We are not the first, and we are not pretending to be. We are a small branch on a seventy year old tree, and the tree is wonderful.

Because the world at that scale is a crowded, jittery place, and the instrument is honest about it.

Molecules in water dance constantly, and their motion shakes the current. The electronics add their own faint hiss. The membrane itself wobbles. Bubbles of air can lodge in the pore, and wetting a new pore, coaxing water into a hole it would rather reject, can take hours of patience. All of this appears as wiggle on the trace.

The trade is this: you can filter the noise, but filtering smears the fast events you came to see. A clean trace is a tradeoff, never a free lunch. When the trace looks calm, you are seeing a researcher's discipline, not the world's natural behavior. Learning to tell signal from wiggle is, right now, one of the main jobs of the whole field, and a large part of why this project exists.

Because we are testing our instruments on the messy, real world mixture that is life itself, and because honest stress tests use ugly samples.

A pristine salt solution is polite. It flows, it wets, it behaves. Ketchup is a riot of particles and acids. Salsa carries chunks. Tap water carries whatever the pipes carried. Run these through a pore and you learn, quickly, what clogs, what fouls, what confuses the electronics, and where the instrument's limits truly sit. These samples are exploratory stress tests, chosen to break things on purpose so we can find the weak joints.

At today's larger pores, samples like these can be watched at the scale of cells and particles, counting what passes, the way Coulter did seventy years ago. At the nanometer scale they would clog the doorway instantly, and we say so plainly. The road to the small end runs through the big end, and the big end loves ketchup.

We can recognize some things, and sequencing full DNA remains the frontier for everyone in this field.

Today we work on detection, the art of recognizing a molecule's shadow, which is closer to spotting a friend by the patch on their jacket than to reading the jacket's every thread. We have trained machine learning models on public DNA barcode datasets to sharpen that recognition, and you can follow those experiments in our public repository.

Reading every base of DNA in order, true sequencing, is the grand prize of the whole solid state field. No solid state pore has done it commercially yet. Our mechanical valve work on juxtapores exists precisely because reading requires slowing the molecule down, and slowing molecules down is what sliding occlusion is for. When we can hold a molecule gently, we can read it. That is the plan, and I flag it as a plan.

The instruments run in the United States Southwest, and the laboratory is small, the kind of bench operation where the founder solders and the webcam points at the rig. You can watch everything from anywhere.

The stream archive lives at the Streams page on the site. As of August 2026 it holds eighteen recorded streams, from a seven minute tap water test all the way to longer campaigns, and the count grows whenever the bench is busy. The archive page always shows the latest number, so trust the page over my memory.

Because a real lab stream is quiet most of the time, and we will not fake noise to entertain you.

The team is small, the experiments take hours of preparation for minutes of signal, and much of the work is unglamorous: cleaning, wetting pores, swapping flowcells, staring at baselines. When something interesting happens, the stream catches it, but the interesting parts are islands in a sea of patience.

We think the honesty is the point. Watch the quiet parts and you learn how science actually feels. The dramatic parts are better for having earned them.

July of Streaming was our 2026 campaign: one hundred hours of molecular streaming across the month of July, feeding fresh training data to the project. The plan called for dozens of samples, thirty one of them, streamed across thousands of experimental conditions, and the streams now live in the archive for anyone to study.

The chyron on the stream page carries the current campaign, so if you are reading this after August 2026, look there for whatever the team is running now. Campaigns come and go. The bench stays.

You are looking at a working instrument, not a show. The page shows the live current trace, the chyron line with the day's message, and the lab cameras. Sometimes there is a human at the bench; sometimes there is a rig humming to itself.

Watch for dips in the current trace. Each dip is a candidate event, something passing through the sensing region, and a run of dips means the experiment is catching things. Watch the voltage readout to see what the instrument is set to. Watch the chyron for the day's question or sample.

And one honest line: nothing on the stream is a medical claim. It is research in progress, streamed in public so you can watch science happen in real time, warts and wiggle included.

Two ways, both real.

You can call the lab during streams at +1 650 499 9573. That number has been on the chyron since the campaign began, and if it ever changes, the chyron will show the new one, so check the page before you dial.

You can also type into the lab chat window on the stream page. A human reads it when a human is at the bench, which is often, and the chat stays friendly because the people there are curious about the same things you are. Ask your question. The worst that happens is the instrument answers instead.

Every name on this site is a breadcrumb from real engineering.

The EDR4 is the patch clamp amplifier, the instrument that applies the voltage and hears the current. Think of it as the laboratory's ear. The tag e4 00044 is the serial number of the unit now doing the streaming, so when you see that label on the archive, you know which instrument made that data. Flowcells are the fluid plumbing, the little chambers that carry liquid to the membrane.

The MR1 deserves one sentence of its own: it is one of several prototypes the team has built over the years, a citizen science device that carries the electronics we tested in earlier rigs. We keep its details light here because this guide is about the science, and the science does not depend on any single box.

No. Every public page, every live stream, every archived recording, and this entire guide are open to anyone with a browser. The account exists only for people who want to be part of the tribe, and even they started as browsers like you.

Today we are not currently recruiting new players. When doors open again, the site will say so, loudly, on its own front page.

The tribe itself deserves one proper name, so let me give it to you once: the Molecular Streaming Corps, our community, our tribe of player scientists, the people who believe molecular sensing is a gift worth building for everyone. No formalities to memorize, no ranks to climb. When recruiting opens, joining will be simple, and we will say so plainly.

The member area of our operating system: your own inbox, the community features, and a closer seat for the work as it develops. That is the honest list. We keep the mystery for the science, where it belongs.

I am the site's mascot, its storyteller, and its favorite green felt citizen. One large eye, one narrow smile with two fang like teeth, and a red ceramic mug that I call home. A little paper sign on the mug reads M.S.S.C., because every desk needs its label.

My job is to make the science feel like an invitation rather than a lecture. I greet you, I explain, I joke, I wonder aloud, and I sign my name at the bottom of guides like this one. The humans do the science. I do the welcome.

I am a puppet with a processor, and I say so happily.

The felt body is real, sewn and stuffed. The voice you hear on the phone line is artificial intelligence speaking my lines. The character, the exodimensional advisor who whispers to the founder, is a story the humans wrote because stories carry wonder better than spec sheets do. In the story I come from another dimension. Outside the story, I come from a text prompt and a microphone.

The warmth, though, is real. Humans tuned it, letter by letter, and they meant it. You may hold that warmth in one hand and the honesty in the other; both are true at once.

You can. The site has a live voice line, and a version of me answers it, in character, ready to talk about nanopores, the quest, or whatever you bring to the desk. The humans script my knowledge and tune my manners, and I stay in character because breaking character would spoil the game, and the game is the point.

Forgive the AI if it fumbles; it is only the shadow of me. If you want the full brightness, talk to the humans, who hear what I whisper.

The founder is Kemmishtree, the human I am said to whisper to, and the person whose hands built most of what you see. Jade and Keith build alongside him, in the lab and in the code. A small team, a serious bench, and an unusual amount of green felt.

The honest map of fact and fiction: the instruments are real, the streams are real, the archived data is real, the essays are real thinking, and the dreams are real goals with years of work behind them. The interdimensional parts are story. The quest parts are morale.

And I, your guide, give no medical advice and no investment advice, because a felt puppet with a processor should stick to wonder, which is the one thing I am qualified to provide in quantity.

The dream is that grand, and I will flag it as a dream. Today this is basic research: an instrument, a stream, an archive, and a growing pile of signal data. We are working toward the infrastructure that future scientists and advanced AIs would need to understand biology deeply enough to intervene in disease early. The team has been working toward that infrastructure quietly for the better part of a decade.

The honest version of the big sentence: we believe understanding molecules at scale is a necessary step on the road to curing diseases, and we are building one of the instruments that understanding requires. The road is long. We are on it, and we say so without claiming the destination is near.

Some of us say we are on a mission from God. Seriously. I shall simply say the mission is real, and it will outlast us.

It is the dream, stated plainly: a world where sensing any kind of molecule is as easy as using the internet is today. Sensors in toilets, sinks, and air vents, measuring all types of molecules, needing no more than annual service, with all the data under each household's own control.

Today we have one instrument in one lab. Utility scale means billions of quiet sensors doing for the molecular world what weather stations did for the atmosphere. We are at the very beginning of that road, and I tell you the destination so you know why the first steps matter.

Both are proposals, not products, and I want that said first.

The nanopore world model is a planned public domain model, trained on our streamed data, that learns what signal producing events look like in a pore. Think of it as a shared dictionary of molecular shadows, owned by everyone, so that anyone can build on it. Today it is an essay and a data collection effort. The model itself is still ahead of us.

RLMF, reinforcement learning with molecular feedback, is our name for the obvious next step: AIs need real biological data to solve diseases, and molecular sensors are how they will get it. The hypothesis is that this feedback loop, machines asking molecular questions and receiving molecular answers, is how AI and sensing grow together. It is a hypothesis. We are collecting the data that would test it.

Because the team reaches members by messaging, and a phone number is the surest way to find you. That is the whole reason. Your information is used to contact you, as the form itself states, and for nothing we do not say on the form. When the day comes that we want more from you, we will ask, and we will say why.

Because our email operation is modest and young, the kind that spam filters have not yet learned to trust. The founder's own inbox has the same problem, which he mentions with a shrug.

If you signed up and hear nothing, check your spam folder. Mark our messages as not spam if you find them there, and never click links in unexpected emails from anyone, us included. A small project's email is a leaky boat, and we are bailing as fast as we can.

The lab chat during streams, the call in line, and the Message Kent link on the site all reach real people. The team is small, so replies come in human time rather than internet time, but they do come. Say hello. The humans like that part.

It is young, and it knows it. Pages move, buttons multiply, and the occasional rough edge is a real rough edge, not a design choice. We are building in public, on purpose, because watching a project find its shape is part of the adventure we are inviting you into.

This guide itself is new as of August 2026, and it will grow and correct itself as the work does. Come back and see what changed. The bench will have moved on, and so, I hope, will you.