The short answer: the real options are auctions, fixed-price marketplaces (SideProjectors, IndieMaker, Microns), brokers (Empire Flippers, FE International), and classifieds. For an app you built with AI, usually pre-revenue and with no clean multiple to anchor a price, an auction fits best: competing buyers set the number instead of you guessing a sticker price a nervous buyer then talks down. That is what ExitBid is built for, a 5-day auction for digital-business exits: AI-built and pre-revenue businesses welcome, a flat $199 listing fee, and 0% commission.
You built an app with AI, with Cursor, Claude Code, Lovable, Bolt, or Replit, it works, and now you want to sell it. The question is not whether anyone will buy it. Most of the 2026 acquisition market is AI-built, and buyers hunt for finished small products to take over. The question is where to list it so the AI origin helps you instead of getting you lowballed.
Two things changed in 2026. The buyer pool now assumes AI assistance by default, and provenance became checkable: Anthropic began embedding machine-readable watermarks in Claude's text output on August 2, 2026 under the EU AI Act's Article 50, and image tools increasingly ship C2PA provenance manifests. That shifts one practical decision most directories skip over, the venue, which this guide answers first, then shows what the new watermarking means for your price.
Related reading
→ How to Sell a Vibe-Coded App in 2026 (Yes, Buyers Pay) → Which Marketplaces Accept Pre-Revenue Businesses? → How to Value a Pre-Revenue Project (When There's No MRR)Where can you sell an AI-built app? The real options
There are four venue types, and for an AI-built, usually pre-revenue app they sort quickly by one question: can the platform price an asset that has no clean multiple? Here are the named places buyers and sellers actually use, with real costs and how each treats AI-built, pre-revenue work.
| Platform | Cost to sell | Format | AI-built / pre-revenue fit |
|---|---|---|---|
| ExitBid | $199 flat, 0% commission | 5-day auction (price discovery) | Built for it: AI-built & pre-revenue welcome |
| SideProjectors | Free to list, 0% | Fixed price / offers | Yes, early and side projects |
| IndieMaker | Listing fee | Fixed price / offers | Yes, with a working demo |
| Microns | 0% (deals under $100K) | Fixed price, curated | Accepts small deals; founder notes buyers pay little for pre-revenue |
| Flippa | $29–$699 listing + flat 10% success fee | Listing or auction | Yes, but geared to revenue |
| Acquire | Free listing; seller closing fee (~6–8%) | Buyer offers / negotiation | Functional SaaS/AI pre-revenue only (exceptions) |
| Brokers (Empire Flippers, FE) | 15% (EF) / 3-5% (FE) + full vetting | Broker-negotiated | Want established revenue |
Read it honestly. SideProjectors and classifieds are great for cheap, low-friction listing if you already have a buyer or a firm price in mind. Flippa brings the largest audience and built-in escrow, but its success fee bites and its buyers expect revenue. Brokers and Acquire skew toward businesses with real trailing numbers. The gap they all leave is the one an AI-built app sits in: a working, pre-revenue asset with no multiple to anchor a price. That is the case an auction is designed for.
How much is an AI-built app worth, and what does it cost to sell?
With little or no revenue there is no MRR multiple, so an AI-built app is priced as an asset: working code, a domain, any users, and the build time it saves the buyer. In practice, polished pre-revenue projects trade from the low hundreds into the low thousands, with early traction or a strong niche pushing higher. The absence of a clean multiple is exactly why a fixed asking price is a guess. For the full framework, read how to value a pre-revenue project, or get a directional number in about 30 seconds from the free valuation calculator (no signup). On cost, the table above is the quick version: free classifieds and 0% marketplaces at one end, percentage success fees and broker cuts at the other, and ExitBid's flat $199 with 0% commission in between, so you know your entire cost before you list.
Why an auction fits an AI-built business specifically
An AI-built, often pre-revenue asset has no clean multiple to anchor on, and provenance just added a fresh axis of buyer uncertainty. Set your price high and buyers, now armed with detectors, assume you're papering over AI shortcuts. Set it low and you've handed a stranger months of your work. Bulow and Klemperer's classic result is that one extra serious bidder tends to beat skilled negotiation with a single buyer. For an asset with no comparables and a new trust variable, competing bids don't refine the price, they create it, each bidder pricing in their own read of your provenance. A transparent seller gets rewarded by the market instead of squeezed by one nervous buyer.
That's the logic behind the ExitBid format: a 5-day auction where verified buyers bid up from a reserve you set. You publish the full provenance story once, every buyer prices it into their bid, and the auction lands the real number in days instead of months of one-on-one haggling. Flat $199, 0% commission, AI-built and pre-revenue businesses welcome. Settlement happens directly between buyer and seller, optionally through Escrow.com. (For the full menu of venues including brokers and marketplaces, see the best places to sell an online business.)
Does being AI-built (and now watermarked) lower your price?
By itself, no. The buyer's real anxiety was never "AI was used." It was "AI was used and nobody checked." Here is what the new provenance signals actually tell a buyer, and what they do not, because the gap is where your price is won or lost.
| What the buyer checks | What a watermark / manifest tells them | What it does not tell them |
|---|---|---|
| Landing page & ad copy | The words were drafted or edited with Claude (text watermark) | Whether a human reviewed, fact-checked, or rewrote them |
| Product screenshots & brand images | The image was AI-generated (C2PA manifest present) | Whether you hold the rights, or it matches the real product |
| Docs, READMEs, changelogs | Documentation was AI-assisted | Whether the code behind it is sound |
The signal confirms AI was involved and answers only half the question. It says nothing about quality, rights, or review. A text watermark is a statistical bias in Claude's word choices, not a stamp on your character; it survives copy-paste but not a rewrite or translation, and a positive hit means Claude may have processed the text, not who wrote it. Closing that other half, in your favor, is what protects your price at any venue.
The two paths: a bigger discount for undisclosed AI the buyer catches themselves, or a near-zero discount for disclosed AI backed by proof of human review. Same business, same code. The only variable is what you document before they look.
Do you have to disclose that you used AI?
Legally it depends on your jurisdiction and what you represent to the buyer. Practically, yes, because non-disclosure is now checkable. Before you list anywhere, assemble this:
- Disclose specifically. "Landing copy and docs were AI-drafted and human-edited. Core logic is hand-written; boilerplate was AI-assisted." Specific disclosure reads as competence, not confession.
- Show the review, not just the claim. Rotate every secret and say so. Attach a dependency audit. Add tests on the core flow. A buyer pays full price for a business they can run in ten minutes and verify.
- Keep a human-authorship trail for what matters. Material published under your real name, on a surface you control, is provenance in the other direction: a track record a detector can't erase.
- Don't over-claim "no AI." That claim is now checkable, and being caught in it poisons the deal. Understating AI use is the new overstating revenue.
Can you just remove the watermark?
Don't. Stripping provenance is technically unreliable, may breach the platform's terms, and if a buyer catches the attempt, it destroys the one thing the sale depends on: trust. Scrubbing the mark to look less AI-built is the same mistake as inflating your MRR. It feels like protecting the deal while quietly setting it up to collapse in diligence. Document the AI use and the human review instead, and pick a venue that prices your transparency fairly.
Frequently asked questions
The real options are auctions, fixed-price marketplaces (SideProjectors, IndieMaker, Microns), brokers (Empire Flippers, FE International), and classifieds. For an app built with AI, which is usually pre-revenue and has no clean revenue multiple to anchor a price, an auction fits best because competing buyers set the number instead of you guessing a sticker price. ExitBid runs a 5-day auction built for this: AI-built and pre-revenue businesses welcome, a flat $199 listing fee, and 0% commission.
With no revenue there is no MRR multiple, so an AI-built app is priced as an asset: working code, a domain, any users, and the build time it saves the buyer. Polished pre-revenue projects commonly trade from the low hundreds into the low thousands, with early traction or a strong niche pushing higher. Because there is no standard multiple, an auction is often the cleanest way to find the real number. See our guide to valuing a pre-revenue project.
No, not by itself. A watermark only confirms AI was involved, which buyers of 2026-wave businesses already assume. What moves your price is whether the AI work was reviewed and disclosed. Undisclosed AI a buyer detects earns a real discount; disclosed AI with proof of human review earns almost none.
Yes. Most of the 2026 acquisition market is AI-built, and buyers don't reject a business for using AI. They reject unverifiable ones: AI-assisted code with no review, no tests, and undisclosed origins. Give a buyer a business they can run in ten minutes with a documented provenance trail and the AI label barely registers.
Legally it depends on your jurisdiction and what you represent, but practically yes, because non-disclosure is now checkable with a detector. Specific disclosure reads as competence. A false "no AI" claim a buyer disproves mid-diligence can collapse the whole deal.
You shouldn't. Stripping provenance is technically unreliable, may breach the platform's terms, and if the buyer catches the attempt it destroys the trust the sale depends on. Document the AI use and the human review around it instead.
The takeaway
Where you sell an AI-built app matters more than the fact that AI built it. Pick a venue that welcomes AI-built and pre-revenue assets and prices them by competition, disclose the AI and the human review in the open, and let the market set your number instead of one nervous buyer talking it down. The watermark only tells a buyer that AI was involved; what you document around it decides whether that costs you anything.
Related reading
→ The Best Places to Sell an Online Business in 2026 → Where to Sell a Pre-Revenue Project in 2026 → Who Buys Pre-Revenue Apps? The Four Buyer TypesPut Your AI-Built Business in Front of Competing Buyers
An auction prices it by demand, not by a sticker you had to guess. Flat $199, zero commission, AI-built and pre-revenue welcome.