I can't believe it is finally here! Micro is deployed to production and is now available for closed beta testing!!! 🚀🚀🚀

What is Micro?

Micro is a platform that harnesses Generative AI to simplify software development. It enables users to build APIs by defining acceptance criteria—essentially test cases—that outline what the software should accomplish. Our deterministic AI approach interprets these criteria to automatically generate functional code, streamlining the development process.

The primary benefit of Micro is its ability to reduce the time and effort traditionally required in coding by automating the translation of tests into working code. This automation not only accelerates development timelines but also enhances accuracy by ensuring that the generated code precisely meets the specified requirements.

Beyond just code generation, Micro incorporates full deployment and hosting capabilities, facilitating a seamless transition from development to production environments. This allows users to interact with their applications in real-time within the platform, adjust features as necessary, and proceed to deployment without the typical delays of traditional development cycles.

Micro addresses several pressing challenges in software development: it lowers the barrier to entry for those without technical expertise wishing to develop applications, speeds up the production process for seasoned developers, and seamlessly integrates coding with deployment, bridging the gap between development and operational use.

By democratizing the development process and integrating essential deployment services, Micro makes software creation more accessible and efficient, appealing to a diverse range of users, from newcomers to experienced developers.

Key Features in the First Release

The initial release of Micro introduces a suite of essential features designed to demonstrate the core capabilities of our platform. It's important to note that this version is our MVP—just the beginning of what we envision for Micro. At Svtoo, we are committed to expanding and refining our platform based on our long-term vision and, crucially, the invaluable feedback we receive from our beta testers. This iterative approach ensures that we continuously enhance the functionality and usability of Micro, adapting and evolving to better meet the needs of our users.

Here is a brief overview of the key features in the first release of Micro:

1. Intuitive User Interface: The platform boasts a user-friendly interface that guides users through the process of defining their API.
2. Define API Schema: Users can define the structure of their API by specifying the expected input and output data.
3. Define Usecases: Users can create acceptance criteria that outline the expected behavior of their API. This is done via a simple, intuitive interface that requires no coding knowledge.
4. Develop API: Micro interprets the defined acceptance criteria to generate functional code that fulfills the specified requirements.
5. Validate API: Users can test their API within the platform before deploying to ensure that it behaves as expected.
6. Deploy API: Once satisfied with the generated code, users can deploy their API to make it publicly accessible. Deployment takes literally one click! This also includes 🔐 API key management.

Demo video

This is the very first demo video of Micro (and for me as a video creator) that I shot while still running the whole thing on my laptop and storing all the data into the local files:

A brief story behind Micro

The concept for Micro began to take shape in ☀️ August 2022, stemming from my musings, "If I'm already writing all these tests, can I generate logic from them?" This led me to the idea of using minuscule ML models as logical blocks that could, in theory, come together to form a complete application. As someone not originally trained in data science, this exploration required me to take up several courses to get a grip on the basics of machine learning.

By ❄️ the beginning of 2023, what started as a side project became serious enough for me to start a company and dedicate myself to it full-time. The initial months were challenging, with ML models not performing as hoped. This prompted a pivot to exploring Large Language Models (LLMs) at a time when they were just beginning to gain attention for their potential. I delved into research papers and experimented extensively, which not only advanced the project but also spawned a series of blog posts on using LLMs for code generation that I shared on my blog.

By ☀️ mid-Summer 2023, I had switched from my initial ML model approach to using LLMs, which led to the first promising prototype of Micro. The results were more consistent and clearly showcased the potential I had envisioned. The next crucial step was transforming this prototype into a robust platform. This phase involved building a comprehensive infrastructure that not only supports dynamic user interactions but also ensures seamless deployment and scalability in production environments.

This journey of curiosity, learning, and adaptation is a story in itself—deserving of a deeper dive, which I plan to cover in upcoming posts focused on the philosophy and detailed story of building Micro.