Spur: AI agents that test websites and mobile apps the way a customer uses them

01

How much revenue does Spur make?

Each row shows the evidence grade and the source numbers from the list at the end of this page. The grades are explained in How Aiolum verifies.

A New York company started by two Yale graduates in 2024 that joined Y Combinator, raised a funding round, grew to 15 people and sells AI testing agents to large online retailers and software companies, with no published revenue or prices
Fact	Value	Grade	How we know	Sources
Enterprise customers, April 2025	more than 30	C	Pulse 2.0, 18 April 2025, on the funding announcement: "in less than a year already have more than 30 enterprise customers using the technology", naming LivingSpaces.com, Norse Atlantic Airways and Wander.com. It does not say how much they pay.	[4]
Founding team	2 co-founders: Sneha Sivakumar (CEO) and Anushka Nijhawan (CTO)	C	Launch post of 20 August 2024: "Anushka and I met our freshman year at Yale." Yale's engineering school, 22 August 2024: "Nijhawan and Sivakumar ... have founded Spur".	[2, 3]
Team today	15 people	C	Y Combinator company page, read 8 November 2026 (team size field). The page also lists open jobs for founding engineers and a founding designer.	[1]
Founded	2024, New York	C	Y Combinator page: founded 2024, New York City. Yale's article says the company "started as a project in a course" and that the founders worked on an earlier user-experience research app in the fall semester of 2023.	[1, 3]
Y Combinator	Summer 2024 batch, US$500,000	C	Batch from the Y Combinator page. The amount is the CEO's description of the standard deal in Yale's article: "they give you $500,000 and they take about 7.5 % of the company".	[1, 3]
Funding round, April 2025	US$4.5 million	C	Pulse 2.0, 18 April 2025: raised "from investors including First Round, Pear VC, Neo, Conviction, Liquid2Ventures, and Predictive Venture Partners", plus angel investors. The Y Combinator page lists a Business Insider article of 17 April 2025 with the same amount; that site blocks ClaudeBot and was not read.	[4, 1]
Prices	not published; sold through a demo	C	Home page and comparison page, read 8 November 2026. /pricing returns a 404 error page.	[5, 7]
Customers named on the home page	Docusign, Abercrombie & Fitch, Bombas, Vuori, Living Spaces, Wondr Health, OneSafe	C	Home page, read 8 November 2026, under "Customers with real applications", each with a case study. The September 2026 launch post also names HelloFresh, Nextdoor, Bilt and Alo Yoga.	[5, 8]
DocuSign case study	green in nine days; 26 defects caught in three months; 80+ tests	C	Spur's case study, undated: Spur "is now the primary validation layer" for DocuSign's self-serve signup, plan changes and checkout, and replaced four automation stacks. The numbers are the company's.	[6]
Native mobile testing	launched 23 September 2026 for iOS and Android	C	Launch post on Y Combinator: "We started with web testing, and today we're launching native app testing across iOS and Android."	[8]
Time from launch to the April 2025 customer count	8 months	calc	Calculated: August 2024 to April 2025.	[2, 4]

How to read this table: a grade C value comes from the company's own pages and launch posts, its Y Combinator profile, or a press write-up that repeats the company's announcement. A "calc" value was computed on this page. No value is grade A or B, and no revenue figure exists.

Spur publishes no prices. The site has no pricing page (/pricing returned a 404 error on 8 November 2026) and sells through "Book a Demo" (grade C)
Plan	Per month, billed yearly (US$)	Per month, billed monthly (US$)	Limit
Annual plan	custom	custom	priced by test-run volume, not per seat; quote after a demo

Spur's comparison page against Mabl says only that "Spur's pricing is built so you run what you need, when you need it", with no figure. At launch in August 2024 the founders offered free setup of tests for a customer's core flows with a launch code, also without a price. The customers named on the home page are large online retailers and software companies.

02

When did Spur launch, and how did it grow?

From a Yale course project to a 15-person company that tests websites and phone apps for large retailers, in about three years. Each entry carries its grade.

Fall 2023CThe two founders, Yale computer-science students, build a user-experience research app in a software engineering course, then apply to Y Combinator at the end of the semester. [3]
Spring 2024CIn their senior year they build a no-code testing tool powered by AI web agents, with a US$500 student grant from Yale's Tsai CITY innovation center, by Yale's article. Both graduate in May with computer-science degrees. [3]
20 August 2024CLaunch post on Y Combinator, "Spur - Your AI QA Engineer": tests written in natural language and run by multimodal agents. The founders offer free setup of core-flow tests at sign-up. [2]
18 April 2025CPulse 2.0 reports a US$4.5 million round from First Round, Pear VC, Neo, Conviction and others, and more than 30 enterprise customers. [4]
23 September 2026CSpur launches native iOS and Android app testing and names HelloFresh, Nextdoor, DocuSign, Bilt, Alo Yoga and Abercrombie as users. [8]
November 2026CThe Y Combinator page lists a team of 15. The home page shows case studies for DocuSign, Abercrombie & Fitch, Bombas and Vuori. [1, 5]

03

Why did Spur work?

Four causes, in the order they mattered. Each one is a reusable lesson.

01

Tests that state the goal, not the clicks

A Spur test says what the user should be able to do, for example finish a checkout with a promo code, and the agent works out the clicks. The founders' argument is that the same test then keeps working across iOS, Android and page redesigns, because the user's goal has not changed.

02

Spur runs the test machines

The customer does not maintain browsers or phones. The home page describes isolated browser sessions started on demand, traffic that leaves from the country being tested, and Spur-owned racks of iPhones, iPads and Android devices. Tests run on each pull request, each release, on a schedule or by an API call.

03

Large retailers and venture money

Spur sells to large online shops and software companies through demos and detailed case studies, and it raised a funding round eight months after launch, as Pulse 2.0 reported. It was built by a funded team that grew to 15 people, not by one person.

How it was built

These details come from the company's launch posts and pages. They are grade C. [2, 5, 6, 8]

The 2024 launch post says Spur uses "multimodal agents", models that read the screen as an image as well as text, and that its tests are not bound to CSS selectors.
Every run keeps console and network logs, video, screenshots and the page structure (DOM). At DocuSign, defects go straight to Jira.
Browsers run as concurrent, isolated sessions; native app tests run on phones in Spur's own device labs, by the home page.
The Model Context Protocol (MCP), a standard that lets AI coding agents call outside tools, is "a first-class interface to Spur", so a coding agent can ask Spur to check a change, by the September 2026 launch post.
The home page states SOC 2 Type II compliance. No source read names the AI model provider.

Tools named in the sources

Each one opens its page in the Tools tab, with every story that uses it.

WebflowB GSAPB CloudflareB TallyC CalendlyC HubSpotB PostHogB Google tagB Google AnalyticsB LoomC

04

Who tried this before Spur, and why did they fail?

Two companies in our database of failed startups promised to take manual quality checking off people's hands and failed. The values below are from that database and are not independently verified.

Died 2012 to 2015Burned US$120K

Testmunk

The promise

Automated mobile app testing on real devices, to replace manual test work.

Why it died

Better-funded competitors and constant operating-system updates that kept breaking the tests.

Competition

Died 2023 to 2026Burned US$7M

Movley

The promise

Quality control for manufacturers through human inspectors, reported in real time.

Why it died

Its human inspectors cost more than AI-based checking, and it did not change in time.

Competition

The failed startups come from our database of 1,749 dead startups, retold here in our own words. Its entries are AI-assisted summaries of public sources, so these cards carry no evidence grade.

What changed

Testmunk sold automated tests on real phones and, like Spur, went through Y Combinator. Its tests broke with each operating-system update and its device farm was expensive. Spur runs its own phone racks too, so that cost is the same; its claim is that a test written as a goal survives interface changes that broke script-based tests. The sources read do not measure how often that holds.

Movley shows the other side: a quality service built on people lost to cheaper automated checking. Spur sells that automation to online shops, whose merchandising teams checked product launches by hand, by its Bombas story. The open question is price competition, because many funded companies now sell AI testing agents.

05

Can one person copy Spur now?

Not as it was done.

Spur was built by two founders with research experience in web agents, Y Combinator money, a later funding round and a team of 15, and it sells to large companies, with a stated security certification and its own device labs. One person could build a narrow testing agent for one kind of shop or one kind of flow, but not this company's breadth.

Where it is still open: Narrow testing agents for one platform or one kind of flow (shop launches, translations, accessibility), where a general testing agent is weaker or too expensive for small teams.

1Pick one check that teams do by hand on every launch, such as product pages and checkout in online shops.
2Let customers write the test as a goal in plain English and let an agent work out the clicks.
3Report evidence with every run: screenshots, logs and a short video.
4Publish customer case studies with numbers, as Spur does for each named customer.

What could go wrong for Spur?

The risks are the mirror image of the causes above.

What is not verified

No revenue figure was found. The headline, more than 30 enterprise customers in April 2025, rests on the company's announcement as Pulse 2.0 reported it and does not say what those customers paid. No payment-processor profile, company filing or interview with revenue was found. The Business Insider article on the round is listed on the Y Combinator page but blocks ClaudeBot and was not read. A reader could look for a revenue figure in a later funding announcement or a founder interview.

A funded team, not a solo business

Two founders launched it in August 2024. It joined Y Combinator, raised US$4.5 million by Pulse 2.0's report of April 2025, and the Y Combinator page lists 15 people in November 2026.

Competition

Spur publishes a comparison page against Mabl, another testing tool, and its DocuSign story describes Playwright tests paused and UiPath kept only for maintenance. On this site, QA.tech is a grade E story about a Swedish company that also sells AI agents that test web and mobile apps like a real user.

Dependence on model providers

The product runs on AI models (the 2024 launch post says "multimodal agents"), but no source read names the provider. Model prices and quality changes affect the cost and reliability of each test run in a way the sources do not quantify.

Company-stated claims

The SOC 2 Type II compliance on the home page and the results in the case studies (for example 26 defects caught at DocuSign) are the company's own statements; no auditor's report or customer statement outside Spur's site was read. The home page's logo strip, "Companies that work with Spur", also shows logos of Sephora, Koala, Boss and True Religion with no case study, quote or other source read; Bilt and Nextdoor are named only in the founders' September 2026 launch post.

More in Developer tools and coding agents

Other businesses in the same group, best evidence first. All 65 stories in the group, 7 with grade-A revenue.

Faster, cheaper serving and post-training of open-weight AI models, sold as a per-token API · United States · Funded

Eigen AI

Evidence A

Sold for $643M

Started small, grew after funding · sold 11 months after launch

Its founders wrote some of the methods that the industry already used to run models cheaply, such as AWQ for 4-bit weights. They turned that research into a paid serving API, proved the speed on a public benchmark, and partnered with a cloud company that needed exactly that speed. Eleven months after its first public post, the partner agreed to buy it.

4 grade-A facts · 4 that failed before it

Paid library of Agent Skills for AI coding agents · China

Skillry

Evidence A

$7.7K in 30 days

Solo founder · 2 months after launch, revenue flat

AI coding agents can build a working page or deck, but in the founder's words their visual output often looks generic. Skillry sells finished visual styles, one task per Skill, shows real output before the sale, and adds new Skills almost every day under one subscription. When sales began on 28 September 2026, they held at about US$430 a day for 17 days.

5 grade-A facts · 4 that failed before it

Paid kits of skills, workflows and subagents for AI coding agents · Vietnam

AgentKit (formerly ClaudeKit)

Evidence A

$82.2K verified revenue

Solo founder · 12 months after launch, revenue falling

It launched in October 2025, when Anthropic's Claude Code had made skills, slash commands and subagents a normal way to steer a coding agent, and when most developers were still writing their own. A founder who presents themself as a CTO with 18 years of experience sold a tested set as a one-time purchase with a money-back promise. Verified revenue reached US$16,040 in March 2026 and US$82,195 in total, but it fell to under US$2,000 a month by September 2026.

6 grade-A facts · 4 that failed before it

Voice-controlled Mac workspace for AI coding agents · USA

CNVS

Evidence A

$50.6K verified, all time

Solo founder · 4 months after launch, revenue falling

Said over US$36K in a monthRecords US$39.6K by end of JulyHeld

It turned a new working habit, running several AI coding agents at the same time, into a native Mac app with voice control and a strong visual style, and it charged once instead of every month. The founder sold it to his own audience through daily live streams, X and a Discord community. Stripe recorded more than three quarters of the revenue in the first seven weeks.

6 grade-A facts · 4 that failed before it

Desktop app that compresses the input of AI coding agents · Netherlands

Headroom

Evidence A

$2K MRR

Solo founder · 6 months after launch, revenue growing

The compression engine already existed as a free, open-source Python tool, but it was hard to install. Garm Lucassen packaged it into a desktop app that installs with a double click, connects the coding agents by itself and removes its changes when it quits. The engine's creator suggested that he build it and keep the Headroom name. Developers who hit the usage limits of their AI plans pay a small monthly price to make those limits last longer.

13 grade-A facts · 4 that failed before it

Email mailboxes for AI agents, by API, MCP or CLI · United Kingdom

Robotomail

Evidence A

$331 MRR

Solo founder · 8 months after launch, revenue flat

AI agents that run for hours or days need an address that people can write back to, and a normal personal mailbox needs a browser login and OAuth consent screens that a program cannot click through. Robotomail sells a mailbox that a program can create and read without a person. It started in March 2026 and connects to Claude, Codex, Cursor and other agents through MCP, an API or a command-line program.

6 grade-A facts · 4 that failed before it

Built with the same tools

Businesses whose sources name tools that Spur also uses: Webflow, Google tag, HubSpot, Google Analytics. Every tool has its own page in Tools.

Trust, estate and probate software for law firms · Ireland, now USA · Funded

EstateMin

No revenue figure

20+ paying law firms

Team of 2 to 10 · 4 months after launch, revenue growing

Probate work in small firms still runs on paper forms, spreadsheets and retyping, and general practice software does not know what an estate file needs. The founders combined a probate legal executive, a former Clio product and customer-success lead and a machine-learning engineer, and they built the product to sit next to Clio instead of replacing it.

Every number graded · 3 that failed before it

AI appointment setter on WhatsApp and website chat · Germany, US company

Setter AI

Evidence A

$104K in 12 months

Team of 2 to 10 · 32 months after launch, revenue falling

It sold one promise that a customer can measure: a first reply to every new lead within 10 seconds, to businesses that already pay for many leads through Meta ads. A demo on the website brought sales calls through organic search before the product was built. Revenue then fell from US$9,000 to US$12,000 a month in late 2025 to under US$2,000 in September 2026, in the same months in which the founder launched other products.

7 grade-A facts · 5 that failed before it

WhatsApp business assistant for tradespeople · Argentina, company registered in Ireland

Tutti

Evidence A

$2.7K in 30 days

Solo founder · 25 months after launch, revenue growing

Tradespeople already use WhatsApp all day on the job site, and many write quotes by hand or in a spreadsheet. Tutti puts the office work into that chat, so the customer has nothing new to install or learn. It sends messages through Meta's official WhatsApp Business Platform, which lets one founder run the service without a phone app as the main product.

5 grade-A facts · 4 that failed before it

AI music-video generator for musicians · Germany

Neural Frames

Evidence C

$1.4M/yr run rate

Team of 2 to 10 · 29 months after launch, revenue growing

It began in January 2023 as an easy web interface for animation with Stable Diffusion, when the alternative was an open-source library that only programmers could run. Three months later the founder rebuilt it for one group of customers: musicians, who need a video for every song they release and cannot pay a film crew. He calls this the best decision he made.

Every number graded · 5 that failed before it

Sources

The grade next to each source is the best grade that the source can support. "Read in full" means that the text of the page was downloaded and the numbers were checked in it.

C Y Combinator: Spur company page
Read in full, 8 November 2026. Source for the founding year, the team size of 15, the Summer 2024 batch, the founders' roles, the news list and the job openings.
C Y Combinator Launch, 20 August 2024: Spur - Your AI QA Engineer
Read in full. The founders' launch post. Source for the founding team, the natural-language tests, the multimodal agents and the launch offer.
C Yale School of Engineering & Applied Science, 22 August 2024: With launch of tech company Spur, recent grads discuss the hectic, exciting startup life
Read in full. Interview with both founders. Source for the course project, the earlier app, the graduation and the Y Combinator deal.
C Pulse 2.0, 18 April 2025: Spur: $4.5 Million Raised To Advance Quality Analysis AI Agent Technology, by Amit Chowdhry
Read in full. A write-up of the company's funding announcement. Source for the round, the investors and the 30 enterprise customers.
C Spur home page
Read in full, 8 November 2026. Source for the product, the test infrastructure, the named customers and the SOC 2 statement.
C Spur case study: Docusign made its revenue path safe to ship with one validation layer
Read in full, 8 November 2026. The company's case study. Source for the DocuSign results and the Jira reporting.
C Spur blog: Why E-Commerce Teams Are Switching from Mabl to Spur
Read the comparison section, 8 November 2026. Source for the pricing statement and the competitor.
C Y Combinator Launch, 23 September 2026: Spur - testing infrastructure for your mobile apps
Read in full. The founders' second launch post. Source for native mobile testing, the MCP interface and the named users.
C Database of failed startups, local export
Local export. AI-written summaries, not independently verified. Source for the failed predecessors.

Run this company? Send corrections or a verified figure to aiolum@livyden.com