Introducing Fractl: Generative AI-powered Programming Language

Authored by Ranga Rao
GitHub repo of Fractl Language

Generative AI is poised to revolutionize the entire software development landscape, with the promise of fluently writing code on behalf of programmers. However, programming is a discipline that requires precision and deep reasoning, not just fluency: minor nuances in code can lead to wildly different outcomes. As a result, AI-driven code generation needs to involve a human (developer) in the loop. Developers still need to meticulously review, refactor, and test AI-generated code (aided by AI, of course) before incorporating it into their projects. Traditional programming languages are a poor fit for this human-in-the-loop workflow: generated code is too low-level, syntactically verbose, and comprehensible only by highly-trained experts, even then requiring significant effort.

Fractl is purpose-built for Generative AI

Today, we are open sourcing Fractl, a new programming language that is designed for Generative AI-code generation. Fractl, with its ultra high-level abstraction and semantics akin to natural languages, is a perfect fit for modeling the business domain of applications. This enables humans-in-the-loop to focus on business logic and harness LLMs effectively to build applications not just with great speed, but also with great precision.

Fractl is built on top of the Clojure programming language and apps built in Fractl can leverage the universe of Java and Clojure libraries.

Fractl is a bridge between Code and Low-code/No-code

With the rising popularity of Low-code/No-code tools, a growing chasm has emerged between traditional programming and visual building. Traditional programming by professional developers offers immense power and flexiblity, but tends to be slow, expensive and constrained by the availability of skilled developers. In contrast, low-code/no-code tools provide speed and simplicity, but are often limited to simple use cases, lead to strong vendor lock-ins, and, as a result, have been strongly rejected by professional developers.

Fractl, with its high-level of abstraction and data-oriented syntax, bridges this chasm by being a single abstraction for both traditional programming and visual building. Code blocks in the Fractl programming language can be directly represented as visual elements in Fractl Design Studio, without any translation, and vice versa. This unique characteristic of Fractl makes it possible for developers to concurrently use multiple ways of building:

  • Traditional coding in IDEs,
  • Visual development in Fractl Design Studio, and,
  • AI-driven code generation in Fractl Design Studio
With Fractl, we finally have a visual development paradigm that is open and works with established developer workflows and best practices.

Fractl is innovative

Fractl innovatively brings together a number of unique programming concepts:

  1. Data-oriented syntax: Fractl code is essentially pure data in Clojure’s EDN encoding. This homoiconic nature acts as the basis for a lot of innovative tooling, and, visual development in Fractl Design Studio.
  2. Graph-based Hierarchical Data Model: Business entities and their relationships naturally form a graph. Such entities and relationships are first-class constructs in Fractl.
  3. Zero-trust Programming: No operations on business entities and relationship are allowed, by default, making apps more secure. Fractl's innovative declarative access-control mechanism, based on the graph structure, brings simplicity to some of the most complex aspects of SaaS development - multi-tenancy, RBAC, and data isolation.
  4. Reactive Computation Model: Business logic is activated in response to data changes or events. This computational model is a natural fit for business processes.
  5. Declarative Dataflow: Express business logic as purely-declarative patterns of data.
  6. Resolvers: Build applications that interact with external systems with a powerful connector mechanism.
  7. Interceptors: Extend the fractl runtime with custom capabilities.
  8. Managed Entity-graph-Database Mapping: Take advantage of an abstract persistence layer for fully-automated storage of entity instances. SaaS platform for building in Fractl

We are announcing the private availability of is our SaaS platform for developers to build and launch applications using the power of Fractl language and Generative AI.

  1. Design Studio: Browser-based development environment for building apps visually and with Generative AI Co-Pilot. Design studio has a native integration with GitHub - software developed in design studio is automatically committed and pushed to GitHub.
  2. Cloud: Managed cloud and DevOps platform for seamlessly deploying apps built in Fractl

Our Story

My co-founder, Vijay Mathew, was a principal engineer at Concur, focused on designing and scaling IAM microservices for the platform. Being a programming language enthusiast, his attention turned to designing a language that would simplify development of planet-scale apps, similar to the one he was working on. Concurrently, I was working on building cross-platform integrations at Cisco and experimenting with higher-levels of abstraction to avoid repetitive software development tasks. Our paths crossed, as part of the Clojure community we were both a part of, and we had a melding of minds. Together we channeled our collective extensive experience building software systems into first defining, then designing and finally implementing the Fractl language.

Pre-seed Funding

We are announcing that we've raised a pre-seed round of funding, led by Westwave Capital. January Capital and Arka Venture Labs co-invested in the round, along with several industry executives. We are thrilled to be partnering with investors with deep focus in Enterprise SaaS technologies, to together shape the fundamental building blocks of next-generation software development.

Learn More

Continue reading

No items found.

Get updates to your inbox.

Learn all our industry secrets
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.