-
SucksLess
Inspiration
Software development is exciting, but it often comes with frustrating, repetitive tasks: unclear requirements, difficult bugs, complicated code, endless reviews, and technical documentation.
We asked ourselves one question:
What if AI could take care of the frustrating parts of development so developers could focus on building amazing things?
That's how SucksLess was born.
Our goal is to make software development faster, smarter, and more accessible through specialized AI agents, automation, and a unified developer experience.
We believe AI shouldn't replace developers. It should give them superpowers.
What it does
SucksLess is an AI-powered developer productivity platform designed to simplify the software development lifecycle.
Our platform brings together four specialized AI agents:
1. Requirements Gate
Analyzes project requirements, identifies ambiguities, detects missing information, and helps developers establish clear implementation objectives before writing code.
2. Debug Triage
Helps developers investigate errors, understand potential causes, and discover actionable solutions instead of spending hours debugging manually.
3. Code Explainer
Breaks down complex code into clear, understandable explanations, helping developers learn, collaborate, and maintain projects more efficiently.
4. Code Review
Examines code quality and provides recommendations for improvements in readability, maintainability, potential vulnerabilities, and best practices.
Through a unified web dashboard, users can navigate these workflows without constantly switching between different development tools.
A key principle of SucksLess is distinguishing AI-generated recommendations from independently verified results, promoting transparency and responsible AI-assisted development.
How we built it
We built SucksLess during the BoxLang + AI Hackathon 2026, organized by Ortus Solutions and ANIA in El Salvador.
Our architecture combines the BoxLang ecosystem, local AI models, web development technologies, and automated testing.
Our technology stack includes:
- BoxLang: Backend logic, application routing, and server-side functionality.
- BoxLang AI: AI integration and specialized agent workflows.
- Ollama: Local execution and management of language models.
- Llama 3.1 8B: Language model used for AI-assisted analysis.
- BoxLang SSR (.bxm): Server-side rendering for the web interface.
- HTML, CSS, and JavaScript: Responsive and interactive user experience.
- JSON APIs: Communication between application components.
- Playwright: Automated browser testing and validation.
- GitHub: Version control, collaborative development, and project integration.
We designed the application around modular components, with each AI agent addressing a different developer challenge.
Our development process focused on connecting the user interface, backend services, AI-powered analysis, and testing workflows into one coherent system.
Rather than relying entirely on model-generated responses, we emphasized human oversight, structured outputs, and the importance of verifiable results.
Challenges we ran into
Building SucksLess within a limited hackathon timeframe required us to overcome several technical and collaborative challenges.
1. Limited development time
We had to prioritize features, organize responsibilities, and develop a working prototype within a strict deadline.
2. Learning the BoxLang ecosystem
Working with BoxLang and BoxLang AI required quickly understanding new development patterns, integrations, and application architecture.
3. Local AI integration
Configuring Ollama and Llama 3.1 8B introduced challenges involving hardware resources, model performance, response times, and communication with backend services.
4. Full-stack integration
Connecting AI workflows, server-side rendered pages, application routes, and the frontend required careful debugging and coordination.
5. GitHub collaboration
Multiple contributors working simultaneously created challenges with branches, pull requests, merging changes, and preserving everyone's progress.
6. Building trustworthy AI workflows
We needed to distinguish between an AI suggesting a solution and a system actually verifying that solution through testing.
These challenges taught us that building with AI is not just about integrating a language model. It requires architecture, testing, teamwork, and thoughtful design.
Accomplishments that we're proud of
We're proud of turning an everyday developer frustration into an ambitious, structured AI-powered platform.
Some of our biggest accomplishments include:
- Designing four specialized AI development workflows within one unified platform.
- Building a web experience with BoxLang server-side rendering.
- Establishing an architecture connecting backend functionality and AI-assisted analysis.
- Incorporating local AI infrastructure through Ollama and Llama 3.1 8B.
- Integrating automated browser testing into our development workflow.
- Collaborating as a team using GitHub under significant time pressure.
- Building a modular foundation designed to support future capabilities.
Above all, we're proud of approaching AI as a tool for solving real problems rather than adding AI just for the sake of it.
We didn't want to create another chatbot. We wanted to build a more useful development experience.
What we learned
This hackathon pushed us to explore technologies, solve unfamiliar problems, and improve how we work as a team.
We learned how to:
- Develop server-side applications using BoxLang.
- Explore intelligent workflows using BoxLang AI.
- Configure and experiment with local language models.
- Design specialized agents around concrete development problems.
- Connect frontend experiences with backend logic.
- Apply automated testing to improve software reliability.
- Coordinate contributions through GitHub.
- Prioritize high-impact functionality under tight deadlines.
Most importantly, we learned that the value of AI comes from how reliably it solves a problem, not simply how intelligently it responds.
A useful developer tool must be understandable, practical, and trustworthy.
What's next for SuckLess
The hackathon is only the beginning of our vision.
Our next steps include:
1. Deeper GitHub integration
Allow developers to analyze repositories, inspect pull requests, and connect AI-assisted reviews directly with existing development workflows.
2. Smarter multi-agent collaboration
Enable specialized agents to coordinate with one another, sharing context across requirements analysis, debugging, explanation, and code review.
3. Automated documentation
Generate useful technical documentation from source code and project structure.
4. Stronger validation
Expand automated testing and improve the visibility of verified results, evidence, and system limitations.
5. Broader AI model support
Explore additional local and cloud-based models, allowing developers to select appropriate options based on performance, cost, privacy, and hardware resources.
6. Better developer experience
Continue refining the dashboard, improving response times, and simplifying interactions with AI-assisted tools.
Our long-term vision is to transform SucksLess into a complete AI-powered developer companion that makes software development more productive, collaborative, and enjoyable.
Less debugging. Less confusion. Less repetitive work. More innovation.
SucksLess — Let AI handle the tedious work. You build what matters.
Developed by Team SucksLess
Diego Novoa | Kevin Fonseca | Felícito Escobar | Beatriz Fuentes
BoxLang + AI Hackathon 2026 — El Salvador
Log in or sign up for Devpost to join the conversation.