Client took over development by vibe coding. What to do?
I’ve worked on a project for one year now, a marketplace web application for one of my clients. It involves a web shop frontend, integration with suppliers, payment platforms, product management, stock syncing, and much more. I built the project from scratch with open source components, guided other developers on the project, and was leading all the technical decisions.Last year I started using LLM’s for cumbersome tasks, and since the beginning of this year I started working with Claude Code to implement features. Still, I always need to think about the implementation, and actively direct and correct the bot. As many of you will know, it can speed up development, but I still need to use my more than ten years of experience as a developer. I thought the project for my client would be safe.This changed when they started developing some of their own, much smaller and simpler projects on an AI-powered no-code platform. They immediately concluded this also must be applied to the marketplace project that will run their core business. I tried to convince them with good arguments that this wouldn’t be a good idea, but failed. Since I am not the owner of the code, and don’t want to be a gatekeeper, I instructed them how to participate in the development with their coding agents.The additions they made to the codebase in only a week are huge, around 10,000 lines of code. To be honest, most of the features they introduced are functional, but the performance of the application has suffered already. What I am most concerned about is the maintainability of the project and how we will get this live. Before, I had a clear mental model of how everything was built, and I added human readable documentation where needed. They still want me to participate in the project and work on the most critical parts of the application, DevOps and other parts they and their coding agents will not succeed in themselves.It seems some people are possessed by the promises of AI-tools, and do not have
FL score
out of 100
Verdict
high confidence
Competition
10
competitors found, emerging market, funded players
Trend
No signal yet
An AI-powered tool to transform messy, unmaintainable 'vibe coded' software into secure, scalable, and polished production-ready applications for mid-market teams.
The pain
The gap
Build angle
Strengths
- Extremely real and severe pain point, explicitly confirmed by market data and user frustrations.
- Rapidly growing market driven by AI adoption, with strong investor interest (Qodo $70M Series B).
- Clear willingness to pay, evidenced by custom service pricing and existing tool subscriptions.
- Specific, addressable complaints that current generic tools don't fully solve.
- High leverage for a solo builder in a cutting-edge field (AI, code analysis).
Risks
- High technical complexity for a solo builder to create a truly effective architectural analysis and refactoring tool.
- Crowded competitive landscape with funded players (Qodo, SonarQube, DeepSource) and established open-source tools.
- Risk of being 'yet another code analysis tool' if the unique angle for AI-generated architectural debt isn't sharp enough.
- Monetization could be challenging against free/cheaper alternatives or highly integrated enterprise solutions.
- Validation needed to confirm mid-market teams prefer a product over human services for deep architectural issues.
Next steps
- Conduct in-depth interviews with Lead Developers/CTOs at mid-market companies that are heavily using AI for coding. Focus on their specific pain points related to architectural quality and maintainability, beyond security or simple code smells.
- Develop a very focused MVP concept for an 'AI Code Architectural Health Report' that could be demonstrated with mock data. Get feedback on the most critical architectural issues and the ideal format for actionable recommendations.
- Research specific open-source tools or academic papers on AI code quality assessment and architectural analysis to evaluate feasibility and potential building blocks for a solo builder. Explore ways to integrate with existing AI coding assistants (e.g., Claude Code, GitHub Copilot) to offer 'pre-commit' architectural analysis.
Fly Labs Method
Is the pain real, is there a gap, is it the right time, can one person build it.
- Problem clarity
- 27
- Solution gap
- 14
- Willingness to pay
- 19
- Buildability
- 12
A promising idea addressing a real and severe pain point in the emerging AI-generated code landscape, but with significant competition and high technical complexity for a solo builder.
Value Equation
Dream outcome and how likely it feels, against the time and effort it costs.
A high-potential market with clear pain and growth, but faces significant competition and high technical hurdles for a solo builder to differentiate and execute effectively.
One-Person Business
Curiosity pull, identity fit, and a path from free value to paid for a solo creator.
A project with a crystal-clear problem for a well-defined audience, but the complexity of the solution itself may hinder a solo builder's ability to create a truly impactful simple product.
Viral Frameworks
Hook strength, shareability, and how cheaply it can be tested.
A viable micro-SaaS with a clear value proposition and reachable audience, but faces significant technical risks and needs more product-specific validation against strong competition.
Builder Lens
Evidence the problem exists, timing, defensibility, and a model that fits on a napkin.
This idea addresses a clear and growing pain with specific users and a strong future fit, but needs a very focused initial product to gain traction against the status quo and emerging competition.
Why this verdict
Five lenses, one composite. How scoring works
The angle
This weekend
- Conduct in-depth interviews with Lead Developers/CTOs at mid-market companies that are heavily using AI for coding. Focus on their specific pain points related to architectural quality and maintainability, beyond security or simple code smells.
- Develop a very focused MVP concept for an 'AI Code Architectural Health Report' that could be demonstrated with mock data. Get feedback on the most critical architectural issues and the ideal format for actionable recommendations.
- Research specific open-source tools or academic papers on AI code quality assessment and architectural analysis to evaluate feasibility and potential building blocks for a solo builder. Explore ways to integrate with existing AI coding assistants (e.g., Claude Code, GitHub Copilot) to offer 'pre-commit' architectural analysis.
Who is already there, emerging market
Qodo specializes in AI agents for code review, testing, and governance, focusing on how code changes affect entire systems, organizational standards, historical context, and risk tolerance, especially for AI-generated code.
Pricing: Contact for pricing (Enterprise-focused)
SonarQube is an open-source platform for continuous inspection of code quality, performing static code analysis to detect bugs, vulnerabilities, and code smells across multiple programming languages.
Pricing: Developer Edition starts at $720 annually. Free tier for cloud-based workflows and a Community (open-source) Edition available.
- CodacyVisit
Codacy automates code reviews and helps maintain coding standards and track technical debt across projects with integrations into Git.
Pricing: Starter plan: $8/month for individual developers with unlimited repositories. Team plan: $24/month for unlimited team members and all features.
DeepSource automatically detects issues and suggests fixes to keep your codebase clean, supporting multiple languages and integrating with IDEs and CI/CD.
Pricing: Transparent pricing model based on active contributors (actual numbers not available without requesting a quote).
Veracode Static Analysis provides enterprise-focused cloud-based SAST (Static Application Security Testing) designed for actionable remediation and compliance.
Pricing: Contact for pricing (Enterprise-focused).
Semgrep is an open-source, DevSecOps-focused, lightweight, and fast static code analysis tool with customizable security rules, great for shift-left security.
Pricing: Sign up for a free trial. Offers paid tiers beyond the free trial.
- VibeCodeRescueVisit
VibeCodeRescue helps non-technical founders, SaaS startups, and established teams transform AI-generated "vibe code" into secure, scalable, and polished software ready for production.
Pricing: Contact for pricing (service-based).
- Ulam LabsVisit
Ulam Labs specializes in rescuing projects stuck in legacy chaos, turning 'vibe-coded' apps into secure, scalable systems with a focus on Python and JavaScript.
Pricing: Free assessment, then custom pricing based on an improvement strategy plan and cleanup sprints.
- Pragmatic CodersVisit
Pragmatic Coders offer professional rescue services for vibe-coded applications, focusing on fixing issues, enhancing UX, and optimizing for commercial use and scalability.
Pricing: Contact for pricing (service-based).
- RailswareVisit
Railsware provides vibe code fixing and refactoring services, transforming chaotic codebases into scalable products with a focus on architecture, modernization, and long-term maintainability.
Pricing: Contact for pricing (service-based).
What they charge
Recent news
- Qodo Raises $70 Million to Revolutionize AI Code Review and Verification | Venture Capital
Venture Capital, March 31, 2026
- Qodo Raises $70M to Accelerate Fight Against Software Slop From OpenClaw and Claude Code - GlobeNewswire
GlobeNewswire, March 30, 2026
- Qodo Raises $70M Series B for AI-Driven Code Verification
AI Business, March 30, 2026
- Qodo Raises $70M to Tackle AI Code Risks - The AI World
The AI World, March 31, 2026
- 'Vibe coding' may offer insight into our AI future - Harvard Gazette
Harvard Gazette, April 1, 2026
Market signals
The market for managing and refining AI-generated code, particularly after 'vibe coding,' is a rapidly growing niche driven by the increased adoption of AI coding tools. Recent funding rounds, such as Qodo's $70 million Series B in March 2026, indicate significant investor interest in solutions that address the quality, security, and governance challenges associated with this new paradigm. The overall market for automated code review tools saw a 173.05% rise in funding in 2025 compared to 2024, with over $1.13 billion in funding in the last 10 years, highlighting a substantial and expanding market for code quality and verification. The ability of AI to generate vast amounts of code quickly has shifted the bottleneck from code generation to code verification and understanding, creating a critical need for tools and services that ensure reliability and maintainability.
What frustrates people
- AI-generated code often lacks architectural structure, making debugging challenging.
- Vibe coding can lead to a codebase that works (sort of) but that nobody, including the AI, fully understands anymore, making changes risky.
- Bugs in AI-generated code can be harder to spot and fix, as developers lose their debugging intuition.
- AI-generated code may lead to poor code quality, maintainability issues, and increased technical debt if not properly reviewed and refined.
- Security risks increase because technical debt grows and authentication risks dramatically increase with unverified AI-generated code.
Semble – Code search for agents that uses 98% fewer tokens than grep
Hey HN! We (Stephan and Thomas) recently open-sourced Semble. We kept running into the same problem while using Claude Code on large codebases: when the agent can't find something directly, it falls back to grep, reading full files or launching subagents. This uses a lot of tokens, and often still misses the relevant code. There are existing tools for this, but they were either too slow to index on demand, needed API keys, or had poor retrieval quality.Semble is our solution for this. It combines static Model2Vec embeddings (using our latest static model: potion-code-16M) with BM25, fused via RRF and reranked with code-aware signals. Everything runs on CPU since there's no transformers involved. On our benchmark of ~1250 query/document pairs across 63 repos and 19 languages, it uses 98% fewer tokens than grep+read and reaches 99% of the retrieval quality of a 137M-parameter code-trained transformer, while being ~200x faster.Main features:- Token-efficient: 98% fewer tokens than grep+read- Fast: ~250ms to index a typical repo on our benchmark, ~1.5ms per query on CPU (very large repos may take longer)- Accurate: 0.854 NDCG@10, 99% of the best transformer setup we tested- MCP server: drop-in for Claude Code, Cursor, Codex, OpenCode- Zero config: no API keys, no GPU, no external servicesInstall in Claude Code with: claude mcp add semble -s user -- uvx --from "semble[mcp]" sembleOr check our README for other installation instructions, benchmarks, and methodology:Semble: https://github.com/MinishLab/sembleBenchmarks: https://github.com/MinishLab/semble/tree/main/benchmarksModel: https://huggingface.co/minishlab/potion-code-16MLet us know if you have any feedback or questions!
AI
Postgres extension for BM25 relevance-ranked full-text search
Last summer we faced a conundrum at my company, Tiger Data, a Postgres cloud vendor whose main business is in timeseries data. We were trying to grow our business towards emerging AI-centric workloads and wanted to provide a state-of-the-art hybrid search stack in Postgres. We'd already built pgvectorscale in house with the goal of scaling semantic search beyond pgvector's main memory limitations. We just needed a scalable ranked keyword search solution too.The problem: core Postgres doesn't provide this; the leading Postgres BM25 extension, ParadeDB, is guarded behind AGPL; developing our own extension appeared daunting. We'd need a small team of sharp engineers and 6-12 months, I figured. And we'd probably still fall short of the performance of a mature system like Parade/Tantivy.Or would we? I'd be experimenting long enough with AI-boosted development at that point to realize that with the latest tools (Claude Code + Opus) and an experienced hand (I've been working in database systems internals for 25 years now), the old time estimates pretty much go out the window.I told our CTO I thought I could solo the project in one quarter. This raised some eyebrows.It did take a little more time than that (two quarters), and we got some real help from the community (amazing!) after open-sourcing the pre-release. But I'm thrilled/exhausted today to share that pg_textsearch v1.0 is freely available via open source (Postgres license), on Tiger Data cloud, and hopefully soon, a hyperscalar near you:https://github.com/timescale/pg_textsearchIn the blog post accompanying the release, I overview the architecture and present benchmark results using MS-MARCO. To my surprise, we were not only able to meet Parade/Tantivy's query performance, but exceed it substantially, measuring a 4.7x advantage on query throughput at scale:https://www.tigerdata.com/blog/pg-textsearch-bm25-fu
AI
Consus (YC F24) – LLM gateway for companies handling sensitive data
Hey Hacker News! I'm Eric, the founder of Consus. We route all AI requests to ITAR restricted and controlled unclassified information (CUI) compliant infrastructure. Consus itself satisfies ITAR restrictions, and we recently completed our FedRAMP equivalency audit against the 410 FedRAMP High controls. FedRAMP High is the U.S. government’s strictest security baseline for cloud services handling sensitive unclassified data. A breach of this data could cause severe or catastrophic harm. Achieving FedRAMP equivalency, or authorization, is required before processing CUI information.Demo: https://youtu.be/6MYxJe6P4oYAerospace and defense is experiencing rapid growth. Unfortunately, most companies are struggling to use AI because of necessary data protections & compliance risk. These companies can move even faster if they could use the AI tools that most other industries take for granted.I spent most of my career prior to Consus in aerospace & defense. I launched national security payloads at SpaceX, and was an early employee at Castelion. I’ve watched first hand how compliance red-tape makes it difficult for companies to adopt new technologies and trust vendors that touch their sensitive data. For example, I worked with a satellite manufacturer that doesn’t allow any AI tools, even for non-sensitive work.Even as AI has become extremely useful over the past 9-12 months, I was surprised by how many aerospace & defense companies refused to touch it. They could see how it could be helpful, but they were scared about model providers training on their data, violating an ITAR rule, or risk losing new business because of a failed cyber assessment. These are all rational fears.There were a lot of ways to solve some/all of these problems, but we focused on building a compliant AI inference layer that can "plug" into any AI tool or workflow. Just as OpenRouter integrates with most tools, Consus can connect to tools like ChatGPT or Clau
AI
Questions about this idea?
FlyBot reads the scoring and gives you a second opinion on “Client took over development by vibe coding. What to do?”.