Cursor vs Claude Code: revenue evidence from 90 shipped apps

Neither tool wins on revenue. We track 44 cases built with Cursor and 49 built with Claude Code, but 15 are the same company on both lists, so the real cohort is 78 distinct businesses. Each side holds exactly 10 third-party-verified revenue figures, and both medians land on $15K/mo. What separates them is not earning power. It is what people build.

Contents

• ·What do the top cases on each side earn?
• ·Where do the cases on each side cluster?
• ·What do the evidence grades say?
• ·Which tool should you pick?
• ·What these numbers cannot tell you
• ·Frequently asked questions

What do the top cases on each side earn?

Both sides top out in the hundreds of thousands per month and both bottom out near zero. Excluding the vendors themselves, which appear on both lists as scale references, the highest verified figure is $500K/mo on the Claude Code side and $440K/mo on the Cursor side. Read the table as a map of business shapes.

Grades, first use: ✅ verified means corroborated beyond the founder's word, usually a dashboard read on camera. 🗣 founder-reported means the operator stated it. 📎 creator-relayed means a third party passed on someone else's number. 🔮 unproven means no figure disclosed, or a target.

Fluently is in that table on purpose: a verified $100 MRR says more about a launch week than an unverified $100K. Across cases publishing a single unambiguous monthly dollar figure — 21 Cursor, 20 Claude Code — the median is $15K/mo on both. Ranges, annual-only totals, lifetime figures and non-dollar currencies are excluded, which is why those counts sit under 44 and 49.

Where do the cases on each side cluster?

Cursor cases cluster around shipped products; Claude Code cases cluster around services and data assets. Both lists carry exactly 15 SaaS businesses. But Cursor has 9 consumer apps to Claude Code's 6, and Claude Code has 9 AI-service businesses to Cursor's 1. That 9-to-1 gap is the sharpest real difference in the data.

The service cases are a different job. AI SEO content services for local businesses runs client retainers of several thousand dollars a month off a $100 Claude subscription. Frey's luxury restroom trailer directory washed 71K rows of dirty data down to 725 verified providers, then reported $273/day plus a single $20K+ lead. An anonymous vibe-coding studio turned one account into 5 client projects and a job offer. None of it is a screen a consumer downloads.

The Cursor cluster is the opposite shape: Gravl at $440K/mo, Prayer Lock at $21K/mo, Stoppr at $12,000/mo from a screen-for-screen clone, Opus Clip at a $215M valuation. Directory sites lean the other way, 5 on the Claude Code side to 3 on Cursor's. Claude Code's cohort is also more solo: 37 of 49, against 27 of 44.

One caveat on that 9-to-1 split, because it matters. One Cursor service case is not a sample; it means our Cursor cohort happened to catch product builders. Read the direction, not the ratio.

What do the evidence grades say?

They say the Cursor list is better supported in aggregate, despite an identical verified count. Cursor breaks down as 10 verified, 26 founder-reported, 4 creator-relayed, 4 unproven. Claude Code is 10 verified, 21 founder-reported, 8 creator-relayed, 10 unproven. Eighteen of 49 sit in the two weakest grades, against 8 of 44.

Our read on why: the Claude Code cohort is younger, and terminal-agent content attracts people selling a method rather than a product. Claude corporate team training is filed as 🔮 because "reportedly quotable at $10K-40K per engagement" is a price, not a receipt. An AI directory site build carries a target of $2K-$10K/mo. A target is not revenue, and we refuse to average it in with dashboards that were shown on screen.

The tier mix says it more quietly: Claude Code carries 26 cases at Tier 2 (replicable) to Cursor's 22, and 8 at Tier 3 (watch the space) to 6. Both lists are dominated by businesses someone could copy. Neither is dominated by businesses someone has proven. For the ranked version across every tool we index, see AI coding tools revenue, ranked in our AI coding tools hub.

Which tool should you pick?

Pick on what you are shipping, not on which list looks richer. If the product is a screen a customer touches, the Cursor cases are the closer analogues. If the value is a pipeline, a dataset or a service you deliver, the Claude Code cases are. The evidence supports no stronger claim.

Building a consumer app you will market on the App Store?

Start from the Cursor cases. Prayer Lock took 3 days to write and 6 months to reach 58K downloads, with all the real work in video format and onboarding. Chart Detector AI runs $56K/mo at roughly a 25% margin after ad spend. Both are distribution stories wearing an app.

Selling a service, an audit, or a directory?

Start from the Claude Code cases, where 9 of 49 are service businesses. The pattern that repeats is structured data nobody else has, or industry knowledge a client cannot get from a prompt. Frey's directory and the local SEO retainer play both sell an outcome, not software.

Not a developer, and validating before you build?

Ignore the tool question for now. Launch Fast reached ~$21.8K/mo at 90 days from a non-technical founder who borrowed someone else's distribution. Subscribr pre-sold 50 lifetime licences for $20K before a line of product existed. Neither outcome turned on the editor.

What these numbers cannot tell you

Three things: whether the tool caused the outcome, what the failure rate is, and how clean the attribution is. We only index businesses that published a number. Every quiet failure built with either tool is invisible here, and in every other roundup you will read.

Attribution is the weakest link. Fifteen of our 78 distinct businesses appear on both lists, which is the honest answer to the versus framing: AEO Service, SiteGPT and Profit AI among them. "Built with Cursor" usually means "mostly Cursor, plus whatever else was open."

The downside is thinly documented too. Our Cursor set holds 1 cautionary tale and the Claude Code set 2, including a men's self-improvement app that verified $6K/mo in its first month and still earned the flag. Three warnings across 78 businesses is not a failure rate. It is the number of people who filmed the bad part.

For the per-tool breakdowns, see apps built with Cursor and apps built with Claude Code. For what each tool is, start with what is Cursor AI or what is Claude Code.

More in AI coding tools.

Frequently asked questions

Is Claude Code better than Cursor for making money?

The evidence supports neither answer. Both cohorts carry exactly 10 third-party-verified revenue figures, and the median across cases with a clean monthly dollar figure is $15K/mo on both sides. The difference is business shape, not size: Claude Code's list skews toward services and directories, Cursor's toward consumer apps and SaaS.

Why do 15 companies appear on both lists?

Because founders use more than one tool and say so. Fifteen of the 78 distinct businesses we index credited both, so we publish the overlap rather than assign each company to one side. Any comparison presenting these as two clean, separate populations is hiding the same overlap.

What does "third-party verified" mean here?

It means the figure was corroborated beyond the founder's own claim, most often a dashboard read on camera or an app store payout screen. Founder-reported means the operator stated it and we could not check it. Creator-relayed means someone passed on another person's number. Unproven means no figure was disclosed, or the figure is a target.

Which side has more solo founders?

Claude Code, at 37 of 49 cases against Cursor's 27 of 44 — roughly three in four versus three in five. We would not read that as proof one tool suits solo operators better. It more likely reflects when each cohort was recorded and which creators cover which tool.