Vibe coding: which startup is ahead?
In our AI code assistant market deck, you will find everything you need to understand the market
SUMMARY
Lovable is ahead in vibe coding today, with Replit close enough to take the lead if its current acceleration holds.
The top two are now effectively tied on annualized revenue, but they are winning different versions of the market. Lovable is stronger at pure prompt-to-app creation for nontechnical users; Replit is stronger once the work starts to resemble conventional software development.
Lovable’s strongest evidence is not its signup count. It is the combination of roughly one million new projects a week and hundreds of millions of monthly visits to the products built on the platform, which shows that at least some of those projects escape the builder and reach real users.
Replit has the broader business and the deeper operating environment. Its cloud IDE, terminal, databases, hosting, collaboration tools, framework support, and enterprise controls make it better suited to applications that need debugging, unusual backends, or long-term maintenance.
Emergent has secured a real third position rather than merely joining the hype. Its reported revenue run rate and application volume are far below the leaders, but already large enough to separate it from Bolt.new and v0 commercially.
The product-quality gap is less flattering than the revenue race. Recent benchmarks still show weak engineering completion, broken changes, incomplete backends, and security scores well below production standards across the category.
Replit leads enterprise deployment today because it offers the strongest combination of private infrastructure, networking controls, identity management, data connections, and developer recovery tools. Lovable is closing that gap, especially for internal tools and simpler web applications.
The moats are also different. Replit owns the broadest platform, Bolt.new owns the most distinctive single technical asset through WebContainers, and Lovable has the strongest usage flywheel around a tightly constrained full-stack workflow.
Security remains the category’s biggest unresolved weakness. The platforms are adding scanners, permissions, private deployment options, and administrative controls, but nontechnical users can still publish software faster than they can learn to test access rules, databases, APIs, and third-party integrations.
The ranking could reverse quickly. Replit can take first place by turning its recent revenue surge into durable enterprise growth, while Lovable must prove that its projects remain active, secure, and paid for after the first burst of building excitement.
This market map, featured in our AI code assistant market deck, highlights top companies and startups in the AI code assistant market
Vibe coding: which startup is actually ahead?
Which vibe coding startups are we really comparing?
The serious vibe coding startup race currently comes down to Lovable, Replit, Emergent, Bolt.new, and Vercel’s v0.
All five let users describe software in natural language, generate working code, and publish an application without setting up a traditional development environment. They still serve different audiences. Lovable and Emergent mainly target people who cannot code. Replit covers everyone from first-time builders to professional developers. Bolt.new focuses on browser-based JavaScript development, while v0 fits naturally into the Next.js and Vercel ecosystem.
We exclude Cursor, Claude Code, Codex, Devin, and similar coding agents because they mainly help developers work inside existing codebases. They compete for some of the same spending, but the product promise differs. A Lovable customer may be trying to avoid hiring a developer altogether; a Claude Code customer usually remains directly involved in the code.
Base44 also sits outside the ranking because Wix has acquired it. Firebase Studio, Figma Make, and other products owned by large technology companies remain relevant competitive threats, although they cannot rank as independent startups. Smaller specialists such as Rork, Create, and Magic Patterns have interesting products, but their disclosed commercial scale remains well below the leading group.
The funding figures below include capital raised by the entire company. This slightly flatters Replit and Vercel in a product-level comparison because both companies existed long before their current vibe coding products. We count only completed financing rounds, so Lovable’s recently reported talks about another large investment are excluded for now.
| Startup | Main product | Main position in vibe coding | Cumulative funding |
|---|---|---|---|
| Replit | Replit Agent | Broad app-building platform with a full browser IDE, hosting, databases, and enterprise controls | About $922M |
| Vercel | v0 | AI app creation connected directly to Next.js, GitHub, and Vercel’s cloud | About $863M at company level |
| Lovable | Lovable | Full-stack web app creation aimed mainly at nontechnical builders | About $553M |
| Emergent | Emergent | Web and mobile app creation for entrepreneurs and small businesses | $230M |
| StackBlitz | Bolt.new | Browser-native JavaScript app creation powered by WebContainers | $135M |
Is there a clear leader in vibe coding today?
Lovable is slightly ahead in pure vibe coding today, although Replit has turned the category into a very close two-company race.
Commercially, the gap has almost disappeared. Lovable reports more than $500 million in annualized revenue, while private-market research firm Sacra estimates that Replit reached roughly $525 million. The difference is about 5%, which is too small to treat as a reliable lead when one figure comes from the company and the other is an outside estimate.
Lovable leads on the activity most closely associated with pure vibe coding. Its users now create roughly one million projects a week, and Lovable-built products attract hundreds of millions of visits each month. Replit has the broader platform: more than 50 million registered users, a complete browser development environment, support for many frameworks, and usage inside 85% of the Fortune 500.
The market structure is fairly clear. Lovable is the leading product for turning a prompt into a conventional web application without much technical involvement. Replit leads once vibe coding expands into professional development, complex backends, enterprise data, and long-term application management.
Emergent occupies third place with a reported annual revenue run rate of about $120 million. That gives it roughly one-quarter of Lovable or Replit’s scale. Bolt.new and v0 remain credible products, but neither discloses enough recent standalone revenue or deployment data to join the top three commercially.
If you want more recent data on this point, please see our latest AI code assistant market report.
As this chart shows, and as featured in our AI code assistant market deck, search interest in AI code assistants has increased significantly
Which vibe coding startup has built the biggest business?
Replit and Lovable are effectively tied on annualized revenue, but Lovable appears to have produced that revenue with less outside capital.
Lovable directly reports its figure, while Replit has declined to disclose a current number and the latest estimate comes from Sacra. Emergent’s figure comes from the company during its latest financing. Bolt’s last firmly reported milestone is now old, and Vercel does not disclose how much of its revenue comes from v0.
A rough capital-efficiency comparison gives Lovable an advantage. Dividing annualized revenue by cumulative funding produces about $0.90 for Lovable, $0.57 for Replit, and $0.52 for Emergent for every dollar raised. The comparison is directional because Replit spent years building infrastructure before vibe coding took off.
| Startup | Latest usable annualized revenue | Cumulative funding | Annualized revenue per $1 raised |
|---|---|---|---|
| Replit | About $525M, estimated | About $922M | About $0.57 |
| Lovable | More than $500M, company-reported | About $553M | More than $0.90 |
| Emergent | About $120M, company-reported | $230M | About $0.52 |
| Bolt.new | $40M at its last firm public milestone | $135M | About $0.30, using stale revenue |
| v0 | Not separately disclosed | $863M at Vercel level | Cannot be calculated |
Which vibe coding startup is growing fastest now?
Replit currently has the strongest revenue acceleration, even though Lovable still has the better claim to leadership in pure prompt-to-app building.
Sacra estimates that Replit grew from about $300 million in annualized revenue at the end of last year to $525 million four months later. That works out to roughly 75% growth and $225 million of additional annualized revenue. Replit’s stated ambition is to reach a $1 billion run rate before the end of the year, which would require another near-doubling from the latest estimate.
Lovable moved from $400 million to more than $500 million between February and June. The percentage increase, around 25%, trails Replit’s estimated pace, but adding $100 million from such a large base still makes Lovable one of the fastest-growing software companies around. Both companies now operate near the half-billion-dollar level, so each new percentage point represents substantial revenue.
Emergent produced the most dramatic initial burst. Its run rate rose from roughly $25 million in December to more than $100 million in February. Growth has since become less explosive, reaching around $120 million during the latest funding round. Emergent remains far smaller than the top two, and its recent increase of roughly 20% suggests that the original launch surge is settling into a more normal pattern.
Recent financing points in the same direction without settling the commercial ranking. Replit closed a $400 million round at a $9 billion valuation. Emergent raised $130 million at $1.5 billion. Lovable is reportedly discussing a deal that could double its previous valuation to $13.2 billion, although those negotiations have not yet produced a completed round. Investors clearly see three major companies forming, but current revenue growth gives Replit the strongest momentum among them.
This chart, featured in our AI code assistant market deck, illustrates yearly VC funding for AI code assistant startups
Which vibe coding startup has real usage rather than impressive signup numbers?
Lovable currently provides the strongest proof that vibe-coded products are reaching people outside the building platform.
Lovable reports more than 50 million projects created, around one million additional projects each week, and an average of 720 million monthly visits across the resulting websites and applications. These figures cover several stages of actual use: projects are being started, published, and visited.
Replit’s audience is larger at more than 50 million users, but that total includes students, hobbyists, traditional programmers, and people who joined before Replit Agent existed. Replit’s revenue growth shows that its old audience is monetizing, although the company publishes less information about how many Agent-built applications remain active after launch.
Emergent had more than six million users and 150,000 paying customers when it passed $100 million in run-rate revenue. Users had created seven million applications at that point, and the total has since risen above 12 million. Bolt and v0 have also attracted millions of users, but their latest public figures reveal less about repeat usage or live deployments.
Which vibe coding platform builds reliable apps today?
No vibe coding platform can yet be trusted to produce a secure, production-ready application without human checking, and Replit currently gives users the best tools for recovering when generation goes wrong.
SWE-WebDev Bench recently tested AI app-building platforms across 68 measures covering requirements, engineering, operations, modifications, and security. None scored above 60% on engineering quality, and none exceeded 65% on security against the benchmark’s 90% target. Common failures included missing requirements, polished interfaces attached to incomplete backends, broken changes after later prompts, and weak infrastructure choices.
A separate end-to-end benchmark found that the best model completed only 58% of test workflows successfully. It also found that agents performing more self-testing generally produced better results.
Replit has the strongest recovery environment. Agent can test applications in a real browser, inspect failures, run terminal commands, edit individual files, access logs, and work across different frameworks. Agent 4 also introduced parallel workstreams and automatic conflict handling.
Lovable is usually smoother during the first build because its technology choices are more constrained. That reduces setup decisions and creates a consistent route through React, TypeScript, hosting, authentication, and data. The limitation appears when a project requires an unusual language, backend, or infrastructure setup.
If you want more recent data on this point, please see our latest AI code assistant market report.
This chart, featured in our AI code assistant market deck, breaks down Anyshpere’s playbook in AI code assistants
Which vibe coding startup is furthest beyond demos and ready for large deployments?
Replit is furthest beyond demos and has the strongest all-round delivery stack, while Vercel gives v0 the deepest web infrastructure underneath its product.
Replit supports many programming languages and frameworks, persistent server processes, managed databases, several deployment methods, private applications, regional hosting, version control, collaboration, and direct terminal access. Its enterprise product adds private networking, single-tenant environments, static outbound addresses, identity controls, and data-warehouse connections. A generated application can remain on Replit as it moves from an experiment into maintained software.
Vercel brings a different kind of scale. v0 connects directly to infrastructure already used for large Next.js applications, global delivery, serverless functions, observability, security, and enterprise deployment. Vercel can even deploy applications inside a customer’s own AWS account. The main unknown is how many important production applications originated inside v0 rather than entering through Vercel’s normal developer workflow.
Lovable comes next for mainstream web applications. Its standard route covers the interface, database, authentication, storage, server functions, payments, hosting, and two-way GitHub work. Lovable applications collectively receive around 720 million visits a month, and the company has signed a multiyear Google Cloud agreement intended to support five times more cloud usage.
Emergent has generated more than 12 million applications and supports web apps, mobile apps, dashboards, e-commerce products, and internal tools. Its customer cases include logistics systems, clinical products, subscription businesses, and operational software replacing several existing tools. The company still provides less evidence than Replit or Vercel on large-scale infrastructure, uptime, and complex enterprise deployments.
Bolt and v0 are mature within narrower workflows. Bolt gives users a JavaScript runtime directly in the browser, while v0 connects generated work to GitHub, databases, pull requests, and Vercel deployment. Both support real development, but Replit currently covers more of the full journey from prompt to long-term operation.
Which vibe coding startup is easiest for nontechnical builders?
Lovable is currently the best default for nontechnical web-app builders, with Emergent close behind for mobile products and small-business workflows.
Around 80% of Lovable builders come from nontechnical backgrounds. Its interface hides most infrastructure decisions, visual editing reduces the need to describe every adjustment in a prompt, and the standard stack handles authentication, databases, hosting, and payments.
Emergent reaches a similar audience. Roughly 70% of its users had no previous coding experience, and nearly 40% were small businesses when the company last released a detailed user breakdown. Its agents ask questions before building and support mobile applications, which gives Emergent an edge for business apps designed primarily for phones.
Replit has lowered its entry barrier through Agent 4, visual design tools, and an infinite canvas. The full IDE, terminal, deployment settings, and effort-based billing still expose more technical complexity than Lovable.
Ease of creation does not guarantee a successful business. A Lovable survey of more than 14,000 users found that over 60% were making no money from their projects, despite widespread plans to build businesses or monetize side projects.
If you want more recent data on this point, please see our latest AI code assistant market report.
This chart, featured in our AI code assistant market deck, illustrates yearly funding for AI code assistant startups
Which vibe coding startup gives developers the most control?
Replit gives developers the most control by a wide margin.
A developer can open the generated files, use the terminal, change frameworks, install packages, inspect the database, review logs, configure deployments, collaborate in real time, and replace sections of generated code manually. Replit Agent now works across arbitrary frameworks rather than forcing every project into a small set of templates.
v0 ranks second for teams already committed to React, Next.js, and Vercel. It can import existing GitHub repositories, create branches, commit changes, open pull requests, connect to databases, and move generated work directly into Vercel’s deployment system. A professional web team can use v0 without abandoning its normal review process.
Bolt gives developers unusual control inside the browser through WebContainers, which run Node.js projects without a traditional remote development machine. Users can install packages, operate a terminal, edit files, and run a backend in the same browser session. Its main boundary is the JavaScript and Node.js ecosystem.
Lovable provides full code ownership and two-way GitHub synchronization, so users can leave the platform or bring developers into the workflow. Its preferred React, TypeScript, and Supabase architecture still offers less freedom than Replit. Emergent also supports GitHub synchronization, although its experience remains more agent-led and less attractive for developers who want to control every implementation detail.
Which vibe coding startup is winning enterprise customers?
Replit is ahead in enterprise vibe coding, although Lovable is closing the gap quickly.
Replit reports usage inside 85% of Fortune 500 companies and has built integrations or partnerships with Databricks, Microsoft, Google Cloud, Okta, Stripe, and other large technology providers. Agent 4 was launched for a mixed audience that included enterprise customers such as Databricks, Snowflake, Stripe, and Atlassian. Replit also offers single-tenant environments, private networking, identity management, data connections, regional controls, and dedicated support.
Lovable says more than half of the Fortune 500 use its product in some form. Recently named users include Workday, Asana, and Nvidia, alongside earlier customers such as Klarna and HubSpot. The phrase “use in some form” can cover anything from a few employees experimenting to a company-wide contract, so we give more weight to Lovable’s named enterprise deployments than to the headline percentage.
The direction still looks favorable for Lovable. It has introduced enterprise identity, security scanning, audit tools, regional controls, managed workspaces, and better support for internal applications. Its latest product update also lets internal tools recognize employees already signed into the company’s Lovable workspace, removing a second login step.
Bolt has made a serious enterprise move through Microsoft Azure, Microsoft 365, and Design System Agents that can use a company’s real components and branding rules. v0 has added enterprise Git workflows, security, managed users, and deployment inside a customer’s own AWS account. Both have strong distribution routes, but neither publishes enough contract or revenue data to challenge Replit’s current position.
This chart, featured in our AI code assistant market deck, compares the main business model options for AI developer tools platforms
Which vibe coding startup offers the best value for money?
Lovable usually offers the best economics for a standard web application, while Replit becomes better value when a project needs unusual backends, deeper debugging, or several technical collaborators.
Every leading platform combines subscriptions with credits, tokens, hosting, model costs, or compute. The final bill therefore depends more on project complexity and failed iterations than on the cheapest advertised plan.
Replit Core currently costs $20 a month and includes $20 of credits. Replit Pro costs $100 and includes $100 of credits, more parallel agents, stronger models, collaboration, and longer database rollback periods. Agent and deployment charges draw from the same credit balance.
Lovable combines building, hosting, its cloud backend, and deployed AI features within one credit system. Its narrower stack can reduce costs for customer portals, directories, dashboards, lightweight SaaS products, and internal CRMs because fewer infrastructure decisions are left to the user.
Bolt offers up to 300,000 free tokens per day and one million per month, although token use rises as a growing codebase must be repeatedly read and modified. v0 also uses model-dependent credits, so heavier models and larger contexts cost more.
The largest savings appear when a generated application replaces labour or several software subscriptions. Lovable has documented a company replacing a roughly $40,000 Salesforce contract with a custom CRM, while Emergent describes a roofing company replacing $1,800 a month of software with one operational system. These are selected success stories, not typical customer results.
If you want more recent data on this point, please see our latest AI code assistant market report.
Which vibe coding startup has the strongest moat?
Replit has the broadest moat, Bolt.new owns the hardest single technical asset, and Lovable has built the strongest usage flywheel.
Replit combines a long-running cloud IDE, tens of millions of accounts, collaborative development, hosting, databases, enterprise controls, educational familiarity, and integrations with major cloud and data platforms. A competitor can copy its prompt box fairly quickly. Recreating the surrounding environment would require years of infrastructure work and distribution.
Bolt’s WebContainers are probably the most distinctive individual technology in the group. They allow a Node.js environment, terminal, package installation, and backend execution to run directly inside the browser. StackBlitz developed that capability before the current vibe coding boom, which makes it harder to reproduce than a standard cloud-hosted agent interface.
Lovable’s advantage comes from the volume and consistency of its core workflow. Tens of millions of projects give the company an enormous stream of examples showing where nontechnical builders stop, retry, publish, or return. Its constrained stack also lets the product team improve a smaller number of recurring patterns instead of supporting every possible architecture.
That flywheel still has limits. Lovable, Replit, Emergent, Bolt, and v0 all rely heavily on foundation models supplied by outside companies. A major improvement from Anthropic, OpenAI, Google, or an open model can quickly narrow product-quality gaps. The durable value sits in distribution, infrastructure, workflow data, customer trust, and switching costs rather than exclusive access to the underlying intelligence.
Vercel owns an influential position around Next.js, deployment, and the professional React ecosystem. Emergent is collecting data from a rapidly growing base of nontechnical and small-business users. Those are credible foundations, although v0 must prove that it is more than a powerful route into Vercel Cloud, and Emergent still needs time to show strong retention across several years.
This chart, featured in our AI code assistant market deck, illustrates how market revenue is distributed across customer segments in the AI code assistant market
Are vibe-coded apps secure enough today?
Vibe-coded apps are still too risky to publish blindly, and none of the leading startups has solved that problem.
A RedAccess investigation found around 380,000 publicly accessible assets created through several vibe coding and hosting platforms. About 5,000 contained sensitive corporate material, including medical information, financial records, internal documents, or customer data. Some exposure came from weak generated applications, while other cases involved users publishing private tools without configuring access controls.
The academic evidence points in the same direction. SWE-WebDev Bench found widespread security and infrastructure failures, with every tested platform remaining below a 65% security score. Research on actual AI-generated commits also found that more than 15% introduced at least one issue, and almost one-quarter of the tracked problems remained in the code later.
Lovable has added scheduled scanning, enterprise controls, security documentation, and low-cost penetration testing through Aikido. Replit provides privacy settings, identity controls, private deployments, single-tenant environments, networking restrictions, and administrative policies. Those improvements reduce risk, but they cannot verify every database permission, API route, and third-party integration.
The largest unresolved problem is the user. Non-coders can now publish software much faster than they can learn to test, debug, and secure it.
Which startup leads each part of the vibe coding market?
Lovable leads the core no-code web-app segment, while Replit owns the broader and more technical end of vibe coding.
The commercial race between Lovable and Replit is narrow, but their strongest positions are quite different. Lovable wins when simplicity, visual quality, and fast deployment matter most. Replit pulls ahead as soon as the buyer needs multiple languages, deeper debugging, existing code, enterprise data, or a development team that wants to take control.
Emergent has established the clearest third position by focusing on nontechnical entrepreneurs, mobile applications, and operational software for small businesses. Bolt and v0 remain specialists with defensible technical ecosystems rather than smaller copies of Lovable.
| Vibe coding segment | Current leader | Closest challenger | Size of the lead |
|---|---|---|---|
| Nontechnical full-stack web apps | Lovable | Emergent | Clear lead in revenue, project activity, and downstream usage |
| Multi-language and backend-heavy apps | Replit | Bolt.new | Wide lead because Replit supports more frameworks, infrastructure, and manual control |
| Enterprise internal tools | Replit | Lovable | Moderate lead, with Lovable gaining through large-company adoption |
| Mobile apps for entrepreneurs and small businesses | Emergent | Replit | Narrower lead because Replit also supports mobile development |
| Browser-native JavaScript development | Bolt.new | Replit | Strong technical specialization through WebContainers |
| Next.js and professional web workflows | v0 | Lovable | Clear lead inside the Vercel and Next.js ecosystem |
| Fast visual web-product creation | Lovable | v0 | Lovable leads on full-stack simplicity; v0 remains stronger for design-led engineering teams |
| Overall commercial scale | Lovable and Replit | Emergent | Top two are nearly tied; Emergent remains around one-quarter of their size |
This chart, featured in our AI code assistant market deck, shows how AI coding assistant technology has evolved over time
Which vibe coding startups are actually ahead?
Lovable is ahead overall in vibe coding today, with Replit close enough to take first place if its current acceleration continues.
We rank Lovable first because this comparison focuses on the central promise of vibe coding: allowing people to create and publish software through conversation without operating like traditional developers. Lovable currently has the strongest combination of nontechnical adoption, rapid project creation, public application usage, simple full-stack deployment, and commercial scale.
Replit’s estimated annualized revenue is now slightly higher, but its business stretches beyond the narrow vibe coding category. Its registered-user base also includes many years of students and developers using the older IDE. Lovable’s activity is more directly tied to prompt-to-app creation, which gives it the edge in this particular ranking.
Lovable has reported more than half a billion dollars in annualized revenue and roughly one million new projects a week. Replit has reached a similar estimated revenue level and serves more than 50 million registered users. The first-place gap is narrow, perhaps closer to 55-45 than 70-30.
Replit can reverse the order by converting its recent growth into durable enterprise revenue, making Agent 4 reliably better on complex projects, and approaching its $1 billion target. Lovable must prove that its projects survive beyond their first months, that customers keep paying after the initial building excitement, and that its security and maintenance tools can keep up with deployment volume.
Emergent ranks third. Its latest revenue run rate is around $120 million, and users have already generated more than 12 million applications. That is an exceptional first year, although the company remains roughly four times smaller than the leaders and has less evidence on enterprise use, infrastructure depth, and long-term retention.
Bolt ranks fourth because WebContainers give it genuine technical differentiation and Microsoft offers a promising enterprise route. v0 ranks fifth because its product is powerful, but its standalone economics remain hidden inside Vercel. The ranking between those two could change easily; neither publishes enough fresh commercial data for a confident gap.
| Rank | Startup | Why it holds this position |
|---|---|---|
| 1 | Lovable | Best overall fit with pure vibe coding, backed by huge project activity, downstream usage, nontechnical adoption, and top-tier commercial scale |
| 2 | Replit | Nearly tied overall and clearly stronger for complex development, enterprise deployment, infrastructure, and developer control |
| 3 | Emergent | Fastest credible new challenger, with strong revenue, mobile positioning, small-business adoption, and more than 12 million applications created |
| 4 | StackBlitz / Bolt.new | Owns valuable browser technology and is moving toward enterprise workflows, but recent revenue and active-deployment data remain thin |
| 5 | Vercel / v0 | Excellent for Next.js and professional web teams, although v0’s independent revenue, retention, and application scale remain undisclosed |
If you want more recent data on this point, please see our latest AI code assistant market report.
OUR METHODOLOGY
This analysis compares the independent startups competing most directly in vibe coding: Lovable, Replit, Emergent, StackBlitz’s Bolt.new, and Vercel’s v0. We exclude developer-first coding agents such as Cursor, Claude Code, Codex, and Devin because their main promise is helping developers work inside codebases rather than letting users create and publish complete applications through conversation.
We break leadership into separate dimensions: commercial scale, growth, real usage, application reliability, deployment readiness, ease of use, developer control, enterprise adoption, value for money, security, and competitive moats. The final ranking reflects the accumulation of evidence across those dimensions rather than any single metric.
We prioritize direct company disclosures, official product documentation, published technical benchmarks, and first-hand deployment information. Independent estimates are used when a company does not disclose a comparable figure, which is why Lovable’s annualized revenue is treated as company-reported while Replit’s latest figure is identified as a Sacra estimate.
Funding totals refer to capital raised by the whole company, not only by the current vibe coding product. That makes the figures for Replit and Vercel less clean in a product-level comparison because both companies built substantial businesses and infrastructure before Replit Agent and v0 became central products.
Registered users are not treated as equivalent to active builders, paying customers, deployed applications, or downstream traffic. We give more weight to project creation, published application usage, customer adoption, recurring revenue, and evidence that generated products remain in use after launch.
For product quality and security, we rely on published benchmarks rather than company demos alone. SWE-WebDev Bench and Vibe Code Bench are used to assess end-to-end engineering performance, modifications, operations, testing, and security, while OWASP provides the broader reference point for common web-application risks.
We exclude incomplete financing discussions, promotional claims without attributable numbers, and standalone valuation headlines that do not add evidence about product adoption or commercial execution. Where data is old, estimated, or reported at company level rather than product level, we say so directly.
Key sources include: Lovable’s official website, Lovable documentation, Replit’s official website, Replit documentation, Vercel’s official website, v0, StackBlitz, Bolt.new, WebContainers documentation, Emergent, SWE-WebDev Bench, the SWE-WebDev Bench paper, the Vibe Code Bench paper, OWASP Top 10, GitHub documentation, Next.js documentation, React documentation, Supabase documentation, and Sacra research.
In our AI code assistant market deck, we identify pain points entrepreneurs should prioritize
Related blog posts
- Who is winning AI coding now?
