A free AI-powered tool that reviews your Y Combinator application before you submit it. Built by Lobster Capital, a VC fund 100% focused on YC companies with $40M+ deployed across 100+ YC startups.

Is YC Roaster really free?

Yes, completely free — no catch, no upsell. Lobster Capital built this to give back to the YC community and help more great companies get accepted.

How does YC Roaster work?

Upload your YC application (PDF or Word) and you'll receive AI-powered feedback shortly after submission, scored across 7 dimensions — including traction, team, market, and scalability. Applications that show strong potential get matched with a YC alumni reviewer for deeper, personalized feedback.

What kind of feedback will I get?

A scored evaluation across 7 dimensions with specific strengths, risks, and suggestions tailored to your application. It's not generic advice — the feedback is based on what's actually in your application.

Who built YC Roaster and why?

YC Roaster is built by Lobster Capital, a venture firm 100% focused on investing in YC companies. With 100+ YC companies in their portfolio, this is their way of helping more founders put their best application forward. Not affiliated with Y Combinator.

How is this different from asking ChatGPT to review my application?

YC Roaster is built by a team that invests exclusively in YC companies. The AI uses a purpose-built framework across 7 specific dimensions based on insights from reviewing hundreds of YC applications. Strong applications also get matched with real YC alumni reviewers — something no general-purpose AI can offer.

When should I submit my application?

As early as possible. The sooner you submit, the more time you have to get feedback, iterate, and strengthen your application before the YC deadline.

Will everyone get a Zoom call?

Everyone gets AI-powered written feedback. If you're matched with a YC alumni reviewer and they see strong potential, they may invite you for a 1-on-1 Zoom session.

Does YC Still Require a Technical Co-Founder Now That AI Writes the Code?

This week Steve Yegge's essay "The Last Technical Interview" hit the front page of Hacker News and started one of those threads founders screenshot and send to each other. His argument, drawn from 35 years of interviewing at Amazon and Google: the traditional technical interview never reliably worked, and AI has finally broken it for good. The future, he says, is watching people do real work, not quizzing them at a whiteboard. The same week, another front-page post asked whether AI is causing "a repeat of frontend's lost decade" by flooding codebases with code nobody fully understands. Put those together and you get the question landing in a lot of YC applicants' inboxes right now: if an AI can write the product, does YC still care whether I'm technical?

The Short Answer

Yes. YC still strongly prefers teams with at least one genuinely technical founder, and that has not changed for the S26 or F26 batches. What has changed is the definition of "technical." YC partners are no longer impressed that you can produce working code, because a non-technical founder with Claude Code or Cursor can now do that too. What they are screening for is judgment: can you decide what to build, tell when the AI's output is wrong, and own the architecture when it breaks at 2 a.m.? If you can do that, you are technical in the sense YC means. If your honest answer is "the agent did it and I'm not sure how," you have a gap to close before you apply.

Why YC Still Wants a Technical Founder

YC's preference for technical founders was never really about typing speed. It was about two things: the ability to build and ship without permission, and the ability to keep iterating after launch when the roadmap is just you, your co-founder, and a list of bugs. Both still matter, maybe more than before.

Think about the canonical YC origin stories. Airbnb had Brian Chesky, a designer, but it also had Nathan Blecharczyk writing the code that made the marketplace actually work. Stripe was the Collison brothers, both deeply technical, shipping an API so clean that developers adopted it before there was a sales team. The pattern YC keeps funding is a team that can turn an idea into a running product over a weekend and then improve it every single day. AI agents make that loop faster, but somebody on the team still has to be the one who understands the loop.

What "Technical" Means in 2026

Yegge's essay is useful here because it names the shift directly. The valuable skill is no longer recalling how to invert a binary tree on a whiteboard. It is orchestrating agents, reviewing their output critically, and decomposing a messy problem into pieces a model can actually execute. That maps almost perfectly onto what a YC partner is trying to figure out in a ten-minute interview: do you have the taste and the rigor to build something real?

The risk the "frontend's lost decade" thread points at is the failure mode YC partners can smell instantly. A founder demos a slick product, then gets asked a basic question about how the auth flow works or why they chose a particular data model, and freezes, because the agent made those decisions and they never looked. That is the new tell. It is the 2026 version of the founder who outsourced their MVP to an agency and couldn't change a button without a contractor.

How to Prove You're Technical in Your YC Application

The application and interview reward specifics. A few that signal real builder judgment in the AI era:

Talk about a decision, not a feature. Instead of "we built X," say "we chose Postgres over a vector DB for the first version because our retrieval volume was tiny and we wanted one fewer moving part." That sentence proves you were driving.
Show the loop. YC loves founders who ship daily. "We deploy several times a day and read every error in production" tells them you own the system, agent-written or not.
Know where the AI was wrong. Founders who can describe a time the model confidently produced broken code, and how they caught it, demonstrate exactly the review judgment Yegge says is now the scarce skill.
Don't hide the stack. If you used coding agents, say so plainly. YC has been clear it does not penalize using AI tools. What it penalizes is not understanding your own product.

Do You Need a Technical Co-Founder, or Can You Be a Solo Technical Founder?

Both get funded. YC has backed plenty of solo founders and plenty of non-technical/technical pairs. The honest framing for 2026: you need at least one person on the team who can own the product end to end, and "the AI does it" is not a person. If you are a non-technical founder, the strongest move is still to find a co-founder who can be that owner. The second-strongest move, and one more realistic than it was two years ago, is to become technical enough yourself that you genuinely understand what you are shipping. Agents lower that bar, but they do not remove it.

Get Your Application Roasted Before You Submit

The fastest way to find your own "freeze" moment is to have someone who has sat on the other side of the table ask you the hard version of "so how does it actually work?" before a YC partner does. YC Roaster connects you with YC alumni who tear into your application and your technical story the way an interviewer will, so the gaps surface in practice instead of on the Zoom that counts. If you are aiming at S26 or F26, it is worth doing before you hit submit.

Frequently Asked Questions

Does YC require a technical co-founder? No, it is not a hard requirement, but YC strongly prefers teams that include at least one founder who can build and own the product. Teams without that capability are at a real disadvantage.

Will YC penalize me for using AI coding agents? No. YC has signaled it does not care that you use agents like Claude Code or Cursor. It cares whether you understand the product they helped you build.

Can a non-technical solo founder get into YC in 2026? It is possible but hard. The stronger paths are finding a technical co-founder or becoming technical enough to own the product yourself, which AI tools now make more achievable.

What's the new "technical" bar in the interview? Judgment: choosing what to build, catching when AI output is wrong, and understanding your architecture well enough to defend your decisions on the spot.