News

We Got 19 Businesses Into the ChatGPT App Store. Here's What We Learned.

Most businesses don't know the ChatGPT App Store exists. You can't struggle to get into something you've never heard of. We've had 200+ rejections and 19 approvals. Here's what we learned.

Naveed Rafi

Founding Team, Noodle Seed

Rehan Ali

Founding Engineer, Noodle Seed

April 2026

8 min read

TL;DR

•Most businesses don't know the ChatGPT App Store exists. That's not a struggle to get in. It's a problem they don't know they have.
•As of April 25, 2026 there are 779 apps in the store. Apple launched with 500 in 2008. The opportunity is hiding in plain sight.
•We've had 200+ rejections. The 19 approvals are the proof that the learnings from those rejections actually work.
•Service businesses with 5 to 50 employees are winning: law firms, agencies, specialists, local experts.
•Most rejections aren't technical. They're broken privacy policies, missing T&Cs, and vague test cases.
•Don't frame your app as a shop. Frame it as a way for people to find you, ask questions, and get in touch. The sale happens outside ChatGPT.
•The window is still open. Being the only business in your category is still possible for most industries.

Bloomberg recently ran a piece with a headline that caught our attention: “OpenAI's ChatGPT App Store Took Aim at Apple, But Results Lag So Far.”

They're not wrong. But they're missing the more interesting story.

The ChatGPT App Store isn't struggling because businesses can't get in. It's struggling because most businesses don't even know it exists. You can't worry about getting into something you've never heard of.

Most haven't started asking the question yet. And the ones who have are getting in almost unopposed.

We've had 200+ rejections. We've also gotten 19 businesses approved. The rejections taught us more than the approvals. The 19 are the proof that the learnings actually work.

We built Noodle Seed to connect businesses to every AI conversation. ChatGPT, Claude, Gemini, and beyond. No engineering team, no months of setup.

1,500+

Businesses signed up

Approved by Noodle Seed

200+

Rejections along the way

~2-3%

Of the whole store (as of Apr 25)

The Store Is Smaller Than You Think

The total number of apps in the ChatGPT App Store is tracked daily at track.appsdiscoverability.com. As of April 25, 2026, there are 779. For context: Apple's App Store launched with 500 apps in 2008. ChatGPT has 900 million weekly users. Their app store has 779 apps.

That number keeps moving. So does our share of it. But as of the time of writing, one platform working with small businesses has placed roughly 2-3% of the whole thing. That's the part that matters.

A law firm in Poland, a copywriting agency in the UK, a healthcare real estate specialist in the US, a social impact brand in Germany, a property appraiser, an art gallery, a speaker, a coaching practice. None of them had engineering teams or AI departments. They just had a business and customers worth reaching.

Not a Marketplace. A Discoverability Layer.

Most people hear “app store” and think of the App Store or Google Play. That framing is wrong and it matters.

The ChatGPT App Store is not a place where people browse and download things. It's a discoverability layer. When someone asks ChatGPT “find me a copywriter who specialises in B2B tech” or “I need a commercial lawyer in Warsaw,” they're not browsing. They're asking. The businesses with an app in the store are the ones that get surfaced in the answer.

That distinction changes who benefits. Across our 19 approvals, the pattern is consistent: service businesses with 5–50 employees are winning, not SaaS companies or consumer brands.

The businesses getting approved fastest, and seeing the most meaningful early engagement, are professional services: consultants, agencies, specialists, local experts. These are businesses whose main challenge is being found. Being one of three law firms in the entire ChatGPT App Store, or the only property appraiser, is an asymmetric advantage that's available right now and won't be for long.

The First 2%

The businesses that are in the store right now are not typical customers. They are early adopters. There is a difference.

Noodle Seed was not a finished product when these businesses signed up. Notifications didn't exist. Several features were half built. The process of taking a business from signup to submitted and live on the ChatGPT App Store was manual on our end. We were doing things by hand that should have been automated. These businesses dealt with all of that.

They didn't mind because they saw something most people don't see yet. They weren't waiting for a polished product. They were waiting for an opportunity that no one else was providing. The quirks of a v1 platform are a reasonable price to pay for being first.

One customer stands out. A business owner in South Korea got on a call with us. He couldn't speak English. We couldn't speak Korean. We spent 30 minutes communicating through Google Translate, working out what his privacy policy needed to say so we could submit his app. By the end of the call I genuinely thought there was no way he was going to follow through. I was wrong. He updated his privacy policy and handled the rest over email.

These businesses are roughly the first 2% who see this opportunity clearly. The other 98% haven't started asking the question yet. Many of them don't know there is a question to ask. After all of Sam Altman's billions of dollars and all of ChatGPT's mainstream adoption, you still cannot book a basic service through it without something like Noodle Seed. That gap is enormous. And right now it is mostly empty.

Five Reasons ChatGPT App Submissions Get Rejected

We've submitted dozens of apps and seen every rejection reason OpenAI gives. The blockers are rarely technical. They're almost always documentation and trust signals.

No privacy policy, or a broken one

The single most common rejection we see. OpenAI validates that your privacy policy URL exists, loads, and is relevant to your app. A placeholder, a 404, or a generic template will get you rejected. It needs to be live, specific to what your app does with user data, and genuinely readable by a human reviewer.

Missing or vague terms of service

Your terms need to be findable and coherent. OpenAI's reviewers check for them. If your website's terms predate your AI app by years and make no mention of how the assistant operates, expect a revision request. A quick addition clarifying the assistant's scope is usually enough.

Inconsistent tool hints and vague test cases

Tool descriptions need to be precise, aligned with real functionality, and clearly scoped. If a tool says it can perform an action but fails or responds ambiguously during testing, it gets flagged. Test cases matter just as much. Submissions with vague or incomplete scenarios fail review because OpenAI reviewers actively test against them. If edge cases aren't covered and expected outputs aren't clear, it signals the app hasn't been properly validated.

Review timelines are front-loaded and unpredictable

First submissions typically take 7–21 days. Resubmissions after revisions can be faster or stretch longer depending on the queue. We've seen 4-day turnarounds; we've seen 6-week ones. The lesson: submit early, treat revision feedback as a checklist, and don't assume a second submission will be quick.

Framing your app as a shop is a red flag

OpenAI reviewers are not approving storefronts. If your submission reads like you're trying to sell things inside ChatGPT, you'll have a harder time. What gets approved is apps that help people discover a business, ask it questions, and get in touch. The actual sale, booking, or transaction happens outside the conversation. A lawyer whose app lets people learn about their services and book a consultation? Clean approval. A retailer whose app description reads like a product catalogue with prices? Much harder. Even if you sell physical products, frame the app around helping people find the right thing and connect with you, not around completing a purchase.

The Window Is Still Open

The store is growing, but from a very small base. Being the first or only business in your category is still possible for most industries. That window exists right now and it won't stay open.

Think about what it means to be one of three law firms in the entire ChatGPT App Store. Or the only property appraiser. Or the only B2B copywriting agency. When a ChatGPT user asks a question in your space, you're not competing with thousands of results. You're one of a handful of options, or the only one.

That dynamic doesn't last forever. The store will grow, categories will fill up, and early movers will have established their presence, their usage history, and their review velocity. The businesses that move now will hold ground that latecomers will have to fight for.

This Is Just ChatGPT

Everything described above applies to one store. But ChatGPT isn't the only AI assistant building an app or connector marketplace. Claude has 353 connectors as of April 25, 2026. Gemini is building its integration layer. Perplexity surfaces businesses in its answers.

The AI assistant landscape is fragmenting in exactly the way mobile did between iOS and Android. Except the surface area is larger, the download friction is zero, and the user intent is higher. People asking an AI assistant for a recommendation are already in buying mode.

The businesses that figure out multi-platform AI presence now, not in two years when it's obvious and crowded, will own a distribution channel that compounds quietly while everyone else is still running Google Ads and wondering why CAC keeps rising.

What to Do Now

You don't need an engineering team. You don't need a technical co-founder. You need:

•A live, accurate privacy policy that covers your AI assistant
•Terms of service that reference how the assistant operates
•A solid knowledge base about your business, services, and the questions customers actually ask
•Someone to build your AI presence across platforms and handle the submission process end to end

That's what Noodle Seed is built for. Not just getting you into the ChatGPT App Store, but connecting your business to every AI conversation. ChatGPT, Claude, Gemini, Perplexity. One platform. But even if you don't use us, the move is the same: get in now, while the store is still small and most categories are wide open.

The businesses that show up early in new distribution channels rarely regret it. The businesses that wait until everyone else has figured it out always do.