Lengthy prototype development cycles that delay market entry and increase costs can be a real headache. At Motius, we continuously refine our development processes to help clients create prototypes quickly and efficiently. When we discovered the BETH stack, we recognized its potential as a shortcut to faster builds, leaner infrastructure, and scalable, production-ready workflows. However, we also recognize that not all parts of BETH fit every scenario. This article describes our experiences with BETH, what worked, what didn’t, and the tech stack we’ve found most beneficial.
Rapid Validation Before Market Entry
Global digitalization has dramatically accelerated technological development. New products and technologies emerge at unprecedented rates. Companies and research institutions therefore face increasing pressure to quickly validate innovations and continuously adapt to changing market conditions. But even small prototypes often involve:
- Multiple repositories (frontend and backend)
- Complex infrastructure (multiple containers for frontend, backend, and database)
This increases overhead and slows down time to market. While low-code tools exist, they often lack flexibility and scalability. On the other hand, fully custom setups consume time and resources.
Why BETH Caught Our Attention
Our team has always been focused on delivering efficient and effective prototypes. We were already moving fast, but we knew there was room for even more speed and simplicity, especially in reducing setup overhead and streamlining our toolchain.
While looking for lightweight, high-performance alternatives, we stumbled upon the BETH Stack, a combination of Bun, Elysia, Turso, and HTMX. We saw the potential for reducing complexity while maintaining flexibility. Our goal: Faster POCs, less complexity, and a more streamlined workflow.
- Bun simplified our tooling: It replaced Node.js, package managers, and test runners with a single, fast runtime. Running scripts and managing dependencies became more streamlined.
- Elysia made backend development much smoother: Its ergonomic API design allowed us to quickly build scalable endpoints without unnecessary boilerplate.
- Turso (serverless SQLite) looked like a fit at first, but understanding that they basically forked sqlite and deploy it on fly.io, we understood that this is not something that massively improves our workflow. We found it more suitable for personal projects rather than customer applications. Additionally, we had concerns about data privacy and control.
- HTMX introduced a new way to think about frontend interactions: We could serve HTML directly from the server, cutting out unnecessary JavaScript complexity. While HTMX is a unique approach to frontend development, we found its mental model somewhat unconventional. Issues like bloated network data and unclear error handling made it less ideal for our production use cases.
Additionally to BETH itself, we explored other database options and toolchain improvements. Built-in functionalities for sqlite inside Bun worked nicely, but especially PGlite proved to be a real gem. The flexible use cases and easy integration with tools like Drizzle reduced friction in our workflow as well as made it easier to focus on building features rather than managing setups. Last but not least Biome replaced other formatting and linting tools. Incredible speed and better output messages helped development a lot. Improving the development experience is an achievement in itself.
5 Key Findings: What Worked Well
We tested BETH by building small applications and analyzing its impact:
- Even Faster POCs are possible: With fewer dependencies and streamlined tools, we can get prototypes running in record time.
- Reduced Overhead: We managed to reduce container sizes from 3 to 2, or even 1 for customers to explore the apps we built.
- Improved Developer Experience: The tooling feels modern, efficient, and enjoyable to use, which contributes to development speed.
- Perfect for CI and Testing: PGlite, a WASM-based PostgreSQL, is a game-changer for running in-memory databases for testing and CI pipelines.
- Drizzle ORM Is a Hidden Gem: We initially thought we’d miss Prisma or Sequelize, but Drizzle’s lightweight and type-safe ORMis a perfect fit.
Considerations for Future Use
While the new stack offers significant advantages, we need to consider several aspects when we adopt it for different projects. Some clients require strict adherence to traditional Node.js environments, which might necessitate fallback solutions for corporate infrastructure compatibility. When it comes to production-ready deployment and long-term scalability, AWS and PostgreSQL remain the preferred choice over serverless SQLite solutions. For edge storage capabilities, exploring PGlite, IndexedDB, and lightweight SQL solutions provides new possibilities for distributed applications.
Bonus: Code generation
Because HTMX didn’t result in the desired speed gains, we also explored code generation tools for scaffolding projects. This led us to screenshot-to-code, a tool that takes images and generates Tailwind + React apps. While the results aren’t perfect, they significantly accelerate the process of bootstrapping POCs, allowing us to focus on refining functionality rather than spending excessive time on initial UI setup.
Proposed Stack: The Best of Both Worlds
How This Enhanced Our Approach:
After experimenting with BETH, we realized we could refine our approach to POCs even further. For projects under 50K, we now consider a BETH-based stack to rapidly validate ideas before larger investments. When projects need production-ready applications, we transition to a more robust infrastructure with AWS, PostgreSQL, and Kubernetes. Throughout this process, we've also streamlined our CI/CD pipelines. We now use PGlite for efficient unit testing and Biome for fast linting and formatting, which creates a more cohesive workflow from concept to deployment.
Real-World Impact: Projects That Would Benefit
Some of our past projects could have benefited significantly from this leaner approach:
- Supply Chain Optimization – Faster POCs would have helped validate logistics processes with minimal setup.
- Real-Time Status Monitoring – Reduced complexity could have meant delivering the project quicker.
- Dynamic Quoting Systems – Lightweight tooling would have provided flexibility in early stages.
- Scientific Research Tools – Faster iteration cycles could have led to better experimental software.
Ready to Accelerate Your Prototype Development?
Exploring the BETH Stack has brought us powerful insights to help businesses to prototype faster. While not every tool in BETH was a perfect fit, the insights we gained have helped us refine our development process. Ultimately, our approach now blends the best of BETH with additional tools like PGlite, Drizzle, and Biome. The result? Faster, leaner prototypes without sacrificing future scalability.
Benefit from our in-depth expertise and speed up your POC development with us! Your innovative ideas deserve rapid execution. We want to help you get from idea to prototype – and then to production – in record time. Get in touch if you’re considering a new tech stack or want to push your POC development to the next level. Let’s find the perfect balance of speed, cost, and scalability for your unique vision.
