As of November 12, 2026, monday.com sells vibe coding to its customers. It also tells outside developers that new Marketplace apps built primarily with no-code platforms or AI-generated vibe code are ineligible.
I call that hypocrisy. The company presents prompt-generated applications as a product benefit while treating the same broad development method as a reason to reject third-party products. The distinction needs a better explanation than who supplies the software.
The two positions are explicit. monday Vibe turns prompts into custom applications. The Marketplace submission guide excludes new apps built primarily through no-code tools or AI-generated vibe code. The planning guide repeats the restriction.
The feature is good when monday sells it
The pitch for Vibe is straightforward: describe the application you need and let AI build it. This is not a reluctant concession that some developers happen to use coding assistants. Prompt-based application generation is the product.
A third-party developer sees a different message at the Marketplace gate. An application can be ineligible because of the way it was built, before a discussion of whether its behavior, security or maintenance meets a defined standard.
That is the double standard I object to. If generated software is acceptable when it sits on infrastructure monday controls, the relevant requirement is the safeguard that infrastructure supplies. Name it. Explain what an outside developer must demonstrate to meet an equivalent bar.
Private apps and a marketplace are different
A custom application used inside an account and a product sold to unrelated customers do carry different responsibilities. Marketplace customers need support, reliable permissions, maintenance and someone accountable when the app breaks. monday's own review process examines engineering, security, privacy, documentation and support. Those are legitimate requirements. Submission and review process
The distinction does not justify a vague proxy for those requirements. An app's development method does not establish that its authorization is wrong, its tests are absent or its developer cannot maintain it. Those are things a review can assess directly.
If the intended rule is that developers must understand and own the code they submit, say that. If a particular no-code runtime cannot satisfy a security requirement, describe the requirement. A blanket category of vibe code leaves developers guessing about the thing the policy actually objects to.
monday also helps AI assistants build apps
The company promotes an apps MCP integration connecting AI coding assistants to its application framework. The announcement describes creating apps, managing features and versions, and handling deployments through an AI-assisted workflow.
So this is not a simple ban on AI-assisted engineering. The company actively supports some of it. The problem is that its published rule does not clearly separate the assistance it encourages from the generated code it rejects.
That ambiguity is concrete. An October developer question describes an app written with an AI assistant, reviewed and tested by its author, and asks whether it is eligible. The inspected public thread displays zero replies. That says nothing about private communications, but the public question remains an excellent demonstration of the rule's practical problem. Developer's eligibility question
Judge the software and the person responsible for it
A product can contain generated code and have a competent owner. A product can be typed entirely by hand and be a disaster. Neither fact exempts it from security, reliability or maintenance requirements.
My criticism is not that monday has an obligation to approve every submission. It is that the company sells the promise of prompt-generated applications while leaving outside developers with an exclusion whose boundary is unclear.
Vibe coding is apparently a feature in monday's product catalogue and a liability in somebody else's submission. Calling that hypocrisy describes the inconsistency visible in the company's own pages. A clear, testable standard for third-party software is the missing explanation.