The YC application asks you to explain, in short and specific answers, what you're building, why you're the team, how far you've gotten, and what you understand that others don't. There's no traction requirement, and about 40% of accepted companies are just an idea. The whole game is being specific, honest, and concise in every box.
YC's written application is deceptively short. The boxes are small, which fools founders into thinking the answers are easy, but small boxes punish vagueness. Every word has to earn its place. This guide walks the application the way the course does: question by question, with what each one is really testing and a framework you can fill in with your own specifics.
The application, question by question
Each row is one question. The middle column is what partners are actually reading for; the right column is a shape you can drop your real answer into. This table mirrors the course's 13-video Written Application module one-to-one.
| The question | What YC is reading for | A framework you can fill in |
|---|---|---|
| Company one-liner | Can you say what you do in one clear line? | What you make + for whom + the outcome. No adjectives, no “revolutionary.” If a smart stranger can repeat it back, it works.Common miss: Buzzwords that could describe ten companies. |
| Company description | Do you understand your own product? | Expand the one-liner into 2–3 sentences: what it does, who uses it, and the concrete problem it removes.Common miss: Describing the vision instead of what exists today. |
| Founding team | Why is this the team to win? | Who does what, how long you’ve known each other, and the specific edge each founder brings to this problem.Common miss: Impressive résumés with no link to the idea. |
| Most impressive accomplishment | Are you someone who makes things happen? | One concrete story of you (or a founder) doing something unusually well, with the specific result, not the title.Common miss: A list of credentials instead of one vivid story. |
| Hacking / resourcefulness | Do you get around obstacles others accept? | A time you bent a system, found a shortcut, or got something you “couldn’t” get, with what you actually did.Common miss: Generic hustle claims with no specific incident. |
| Progress | How far have you gotten? | What exists now: built, launched, users, revenue, whatever is true. Idea-stage is fine; be exact about where you are.Common miss: Inflating progress; partners can tell. |
| Traction | Is it working, in numbers? | Your single best metric, its growth, and the window: “X now, up from Y a month ago.” Then how you know it’s real.Common miss: Vanity metrics with no denominator. |
| Product / founder fit | Why you for this problem? | The unfair advantage, personal stake, or years of context that make you the right people to build this specifically.Common miss: “We’re passionate about it.” Everyone says that. |
| Competitors / unique insight | Do you see clearly, and know something others don’t? | Name the honest alternative (often the status quo), then your unique insight: the belief that shapes your product.Common miss: Saying you have no competitors. |
| Monetization / vision | How does this make money, and get big? | Who pays, how much, how often, then the one-sentence version of how a small wedge becomes a large company.Common miss: A giant TAM with no path from step one. |
| Go-to-market | How will you actually get users? | Your first concrete channel and why it fits your users: one channel you can prove, not five you’re hoping for.Common miss: “Viral growth” as a plan. |
| Other ideas | How do you think? | One or two honest alternatives you considered and why you chose this. It shows range, not disloyalty to the idea.Common miss: Leaving it blank or writing “none.” |
Draft written? Get it read back before you submit.
The Grader reads each answer for specificity, evidence, clarity and concision, then shows what to fix.
The three qualities every strong answer shares
Across all twelve questions, the answers that work share the same shape, and it is what the Grader measures:
- Specific. Numbers, names, dates, concrete nouns. "We grew fast" is noise; "we went from 20 to 210 weekly active users in six weeks" is signal.
- Evidenced. Claims backed by something real: a metric, a launched product, a user quote, a personal track record. Assertion without evidence reads as hope.
- Concise. The point, made fast, with the fluff cut. Partners read thousands of these; the ones that respect their time stand out.
Inside the lessons: what strong answers actually look like
The table above gives each answer its shape. The course's thirteen-lesson Written Application module goes further, with real answers from admitted applications, including the instructors' own, read and annotated line by line. A sample of what the lessons cover:
How YC actually reads your application
In the module's opening lesson, YC's head of admissions describes her three reading passes: understand the idea quickly ("you should make it easy for me to have a clear picture of what you do, ideally in the first three sentences"), then understand why this is the right team, then understand what you've done so far. Her editing rule is quoted directly: "Please don't use more words when you could use fewer." Every answer you write serves one of those three reads.
The one-liner: about 50 characters, plainness wins
YC caps the one-liner at roughly 50 characters. Real one-liners from recently admitted companies read like these: "Instant product photography with AI." "Telehealth rehab for lung disease patients." "DoorDash for pharmacies." No jargon, no sales pitch, just what the company does. When an "X for Y" framing helps, YC's own three-part test applies: X must be genuinely successful (not just known), Y must clearly want X's model, and Y must be a huge market. The counterexample the course cites from a YC partner: "Buffer for Snapchat" fails the test because the X is far smaller than the ambition needs.
Founding team: the three questions being de-risked
Accelerators bet on people more than ideas, and the team answer is read against three questions in order. Can this team build the product themselves? Most top programs screen explicitly for a technical founder, and "we'll use the funding to hire an engineer" rarely survives. Will this team stay together? Applications ask how you met because co-founder breakups kill more early companies than competition does; co-founders who started together but hold very unequal equity read as a conflict waiting to happen. Are these founders in it for the long haul? The lesson walks admitted answers that pass all three, including one that establishes a 14-year co-founder friendship in two sentences, alongside the Airbnb founding story as the canonical case of determination outweighing a widely mocked idea.
"Hacked a non-computer system": the formidability probe
YC asks for a time you successfully hacked a non-computer system to your advantage. The lesson reads two real answers. One is an instructor's own: a college-era loophole in Amazon's student-referral program, built on memes and pre-emptively created Facebook groups, that generated about $80,000 a year in commissions at 18 and ended in Amazon hiring him. The other is an admitted founder who, blocked from a real-estate license for lacking a degree, earned an accredited bachelor's in four weeks through a competency-based online university. The question screens for what Paul Graham calls being formidable: "roughly, justifiably confident."
Progress and traction are read against time
$100k of revenue three months in, from two part-time founders, is remarkable; the same figure after five years full-time is a red flag. Partners read for pace. In the words of one founder quoted in the course, the reaction you want is "how did they get this done in that time frame," not "I probably could have done that too." The traction lesson also covers evidence that isn't revenue: GitHub stars for open-source projects, pre-launch landing-page signups, and letters of intent from recognizable companies, plus the theme every accelerator leader repeated to the instructors: how often you talk to real users is itself a signal.
Monetization: size the market from the bottom up
"If we capture 1% of this $2 trillion market…" is the answer partners skim past, because top-down math is detached from any action you can take. The lesson teaches the bottoms-up alternative through a worked example (Thumbtack's model, priced plumber by plumber, city by city, up to a credible national figure) and quotes Sam Altman's version of the ideal answer: a path to covering your expenses within a year, plus a believable sketch of how this becomes a billion-dollar-a-year company in ten. Concrete first customer, then second, then the scale story, in that order.
Go-to-market: unfair distribution is a moat
Dozens of teams pitch every accelerator the same idea, so the go-to-market answer is often the tiebreaker, and "Facebook ads, influencer marketing, cold outreach" is what everyone else writes. The instructors' own admitted answer instead named a specific community they already had privileged access to, plus a tool-built list of 10,000 identified prospects. The principle the lesson draws out: unfair access to distribution is a form of secret sauce, and naming yours specifically beats listing channels anyone could buy.
"Other ideas" is not a trick question
Founders often leave the other-ideas box empty, afraid it signals weak conviction. The lesson explains what it actually gauges: whether you're receptive to guidance. YC sometimes issues conditional acceptances contingent on working on a different idea, so listing honest alternatives generally helps, unless this idea is genuinely your life's mission, which is also a fine answer.
Competitors: teach the reader something
Claiming "no competitors" is the classic red flag. It usually means the status quo is your competitor and you haven't named it. The strongest answers teach the reader something about the market, and the lesson's centerpiece is Cruise's admitted application: "the best way to differentiate a car is with new technology, not leather seats." Four repeatable ways to ground a unique insight follow: an unmet customer need incumbents ignore, a technological cost shift, why-now timing (Uber was only viable once GPS smartphones were ubiquitous), and regulatory change (Ro and Hims followed Viagra's 2017 patent expiry).
What this page can't do is show you the full answers. The lessons read them end to end, including the instructors' own product-founder-fit answer with the two changes they would make to it today. That kind of annotated hindsight only exists in the free course, in Module 3.
How to draft: a working process
Strong applications are edited into existence, not written in one pass. A process that consistently works:
- Brain-dump first. Answer every question badly in one sitting. Long, messy, honest. The goal is to get every true fact about your company onto the page where you can see it.
- Find the lead. In each messy answer, one sentence is the real answer. It is usually buried in the middle. Move it to the front.
- Cut to the limit, then keep cutting. Character limits are a floor for editing, not a target for filling. An answer at 60% of the limit that says everything reads better than one that fills the box.
- Read it as a stranger. A day later, read the whole application top to bottom in one go. Anywhere you skim is somewhere a partner will skim too.
- Get one outside read. Someone smart who doesn't know your company should be able to repeat back what you do, who it's for, and why you. If they can't, the writing failed, not the reader.
Consistency: the check most founders skip
Partners read the application as one document, and inconsistencies read as either carelessness or spin. Before submitting, check the answers against each other. Does the traction number in one box match the progress story in another? Does the go-to-market channel actually reach the user named in your one-liner? Does the founding-team answer explain why these people are suited to the unique insight you claimed three boxes earlier? A single pass for internal consistency catches the contradictions that quietly sink otherwise strong applications.
If you're idea-stage or pre-traction
Do not pad. About 40% of funded companies are accepted at the idea stage, so being early is a normal state, not a defect to hide. What changes is where your evidence comes from: with no metrics to cite, the accomplishment, resourcefulness, and founder-fit questions carry the application. Make those answers vivid and concrete. A specific story of doing something difficult beats a hypothetical roadmap every time. And state your stage plainly in the progress box; "we start building next month, here is what we have validated so far" is a strong answer when it's true.
Idea-stage teams also have one structural advantage worth using: nothing is locked in yet, so the unique-insight answer can carry the whole application. Spend disproportionate time there. A genuinely sharp observation about your market, stated in one plain sentence, is the strongest signal an idea-stage application can send.
Don't forget the founder video
YC also asks for a one-minute founder video and a product demo, and the most common mistake is startlingly simple: not following the directions. A YC partner quoted in the course describes receiving "20-minute spirit-journey videos" and clips missing a co-founder, and notes that an absent co-founder in the video "speaks volumes," sometimes signaling the co-founder doesn't know the application exists. The bar is unglamorous: about a minute long, every founder on camera, each saying who they are and what they do.
The course's videos break down real admitted videos, including DoorDash's, where four founders compress who they are, 100+ customer interviews, a live product, and $10,000 of first-month sales into under a minute, against the instructors' own YC video, with the three mistakes they'd fix in it today. The working method the lessons teach: pick three to five "vertebrae," the core points the video must land (team, traction, unique insight), and build the sixty seconds around those. Personality matters more than production: the videos that work read as founders having fun, not founders reciting a script.
And whatever you do, don't leave the product demo blank. YC's CEO has called the demo his favorite and most important part of an application, and YC's head of admissions, quoted in the course, says unpolished is fine: running locally is fine, "just a CSV" is fine, because "the difference between nothing and something is actually huge." The instructors were accepted with only a Figma prototype on record.
Module 4 · How to Make a Winning YC Application Video · “Application video”
The course's Video Introduction & Demo module plays the real admitted videos, DoorDash's and the instructors' own with mistakes annotated, and covers the demo that YC's CEO calls the most important part of the application.
After you submit
Once the application is in, the process is out of your hands for a while, and knowing the shape of what follows keeps you from refreshing your inbox unproductively. If YC is interested, the next step is an interview invitation; YC gives feedback only to companies invited to interview, so a silent rejection carries no information about why.ycombinator.com/faq Decisions land on the batch's published timeline, and funding begins as soon as a company is accepted rather than waiting for the batch to start.ycombinator.com/apply
The productive move during the wait is the same as the productive move after a "no": keep building. If an interview comes, fresh progress since submission is the strongest thing you can open with. If it doesn't, that progress becomes the backbone of a reapplication, which about half of every batch made before getting in.
Put it together
Write the whole application in one sitting, then walk away for a day, then cut it by a third. Run it through the Grader for a second opinion on each answer, and read your responses aloud; anything that's hard to say out loud is usually hard to read. When the writing is tight, your interview answers get easier, because you've already done the thinking. And keep an eye on the deadline so you're submitting early, not scrambling.