Codev lets enterprises avoid vibe coding hangovers with a team of agents that generate and document code
For many software developers using generative AI, Vibe coding is a double-edged sword.
The process produces rapid prototypes, but often leaves a trail of brittle, undocumented code that creates significant technical debt.
A new open source platform, Codevaddresses this by proposing a fundamental change: treating the natural language conversation with an AI as part of the actual source code.
Codev is based on SP(IDE)R, a framework designed to turn vibe-coding conversations into structured, versioned, and auditable assets that become part of the code repository.
What is Codev?
At its core, Codev is a methodology that treats the natural language context as an integral part of the development lifecycle, as opposed to a disposable artifact as is the case with vanilla vibe coding.
According to co-founder Waleed Kadous, the goal is to invert the typical engineering workflow.
“An important principle of Codev is that documents like the specification Are the actual code of the system,” he told VentureBeat. “It’s almost as if natural language is being compiled into Typescript by our agents.”
This approach avoids the common pitfall of creating documentation after the fact, or not at all.
The flagship protocol, SP(IDE)R, provides a lightweight but formal structure for building software. The process starts with Please specifywhere a human and multiple AI agents work together to convert a high-level request into concrete acceptance criteria. Then in the Plan In this phase, an AI proposes a phased implementation, which is reassessed.
For each phase the AI performs a IDE loop: It Implements the code, Defends it against bugs and regression with extensive testing, and Evaluates the result according to the specification. The last step is Judgementwhere the team documents lessons learned to update and improve the SP(IDE)R protocol itself for future projects.
The key differentiator of the framework is the use of multiple agents and explicit human review at different stages. Kadous notes that each agent brings unique strengths to the assessment process.
“Gemini does extremely good at finding security vulnerabilities,” he said, citing a critical flaw in cross-site scripting (XSS) and another bug that “could have shared an OpenAI API key with the customer, which could have cost thousands of dollars.”
Meanwhile, “GPT-5 is very good at understanding how to simplify a design.” This structured review, with a human giving final approval at each stage, prevents the kind of runaway automation that leads to flawed code.
The platform’s AI-native philosophy extends to its installation. There is no complicated installer; instead, a user instructs their AI agent to apply the Codev GitHub repository to set up the project. The developers “dogfooded” their framework and used Codev to build Codev.
“The key point here is that natural language is now executable, where the agent is the interpreter,” Kadous said. “This is great because it means it’s not a ‘blind’ integration of Codev; the agent gets to choose the best way to integrate it and make decisions intelligently.”
Codev Case Study
To test the effectiveness of the framework, the creators conducted a direct comparison between vanilla vibe-coding and Codev. They gave Claude Opus 4.1 a request to build a modern web-based todo manager. The first attempt used a conversational, vibe-coding approach. The result was a plausible-looking demo. However, an automated analysis conducted by three independent AI agents found that it had implemented 0% of the required functionality, contained no tests, and lacked a database or API.
The second attempt used the same AI model and prompt, but applied the SP(IDE)R protocol. This time the AI produced a production-ready application with 32 source files, 100% of the specified functionality, five test suites, an SQLite database, and a full RESTful API.
During this process, the human developers reported that they never directly edited a single line of source code. Although this was a single experiment, Kadous estimates that the impact is significant.
“Subjectively, I feel like I’m about three times as productive with Codev as I am without it,” he says. The quality also speaks for itself. “I used LLMs as a judge, and one of them described the results as what a well-oiled engineering team would produce. That was exactly what I was aiming for.”
While the process is powerful, it redefines the role of the developer from a hands-on coder to a systems architect and reviewer. According to Kadous, the initial specification and planning phases can each take between 45 minutes and two hours of focused collaboration.
This contrasts with the impression given by many vibe coding platforms, where a single prompt and a few minutes of processing get you a fully functional and scalable application.
“All the value I add is in the background knowledge I apply to the specifications and plans,” he explains. He emphasizes that the framework is designed to augment, not replace, experienced talent. “The people who will do the best… are senior engineers and above, because they know the pitfalls… You just take the senior engineer you already have and make them much more productive.”
A future of collaboration between humans and AI
Frameworks like Codev signal a shift where the primary creative act of software development is shifting from writing code to creating accurate, machine-readable specifications and plans. For business teams, this means AI-generated code can become auditable, maintainable, and reliable. By capturing the entire development conversation in version control and enforcing it with CI, the process turns ephemeral chats into long-lasting technical assets.
Codev envisions a future where AI acts not as a chaotic assistant, but as a disciplined employee in a structured, human-led workflow.
However, Kadous recognizes that this shift brings new challenges for the workforce. “Senior engineers who reject AI outright will be outpaced by senior engineers who embrace it,” he predicts. He also expresses concern about junior developers who may not get the opportunity “to build their architectural talents,” a skill that becomes even more important when mentoring AI.
This highlights a central challenge for the sector: ensuring that AI elevates top performers, but also creates pathways to develop the next generation of talent.
