Huggingface like platform for Programable Prompts
Open Source Cloud based IDE to build, train and monetise cross LLM and high accuracy Prompt Packages powered by Micro LLMs
Try our Prompt Package IDE here : Sugar Factory
Sugarcane AI is creating an open marketplace for developers to package and sell high-accuracy, npm-like programmable prompt packages.
The product is designed to address the issues of prompt accuracy and developer productivity by utilizing a microservices approach for prompt and LLM management, saving 80% of development time and server cost for LLM applications.
Focused on the creation of reusable, high-accuracy prompt packages and Micro LLMs, we offer an open-source microservices solution to empower the next 10 million LLM application developers to build prompt packages and train Micro LLMs.
We believe that a community driven package ecosystem is a long term solution for LLM application development similar to other programming languages ecosystems. 🎉
- Build and ship your prompts over APIs as packages
- Log, Label and Manage prompt completion datasets
- Version and backtest prompts for better performance
- Experiment and test Prompts with multiple available LLMs
No Code UI tool to create and deploy Prompt pacakges as a real-time and shareable LLM based Applications
- Create and share re-usable Prompts in just few clicks
- Collect Feedaback & Training data from actual prompt usage
- Single Click Deploy and Use as an API over any Platform
Demo Sugar Cubes :
Text Generator : Birthday Wish Generator
Image Generator : Santa with Celebrity Face
View Demo : How to Create Sugar Cubes ?
- Publish, share and monetise your prompt packages
- Discover and use managed prompts in Sugar Bakery
- Leverage Open datasets and Finetuned Micro LLMs
Prompt Package Marketplace : Access Marketplace
Prompt Packages are a fundamental component of Sugarcane AI, designed to streamline and simplify AI app development for developers and data scientists. These packages offer a structured, reusable set of prompts tailored to specific tasks or domains, enhancing efficiency and productivity in creating accurate AI applications.
Prompt templates in Sugarcane AI are powerful tools designed to streamline the creation of AI prompts. These templates offer pre-defined structures, ready-made prompts, and placeholders that allow developers and data scientists to craft prompts quickly and efficiently.
Micro LLMs, short for Micro Language Models, serve as the core foundation of Sugarcane AI's microservices architecture. These precision-tuned language models are designed to deliver exceptional accuracy, cost-efficiency, and speed, particularly in app development scenarios. In comparison to conventional base LLMs, Micro LLMs excel by providing developers with superior results while requiring fewer computational resources.
git clone https://github.com/sugarcane-ai/sugarcane-ai.git
cd sugarcane-ai
cp apps/factory/.env.example apps/factory/.env
Replace the values in .env with your own
pnpm install- Run
pnpm devto start service
Explore our Documentation for detailed instructions on getting started with Sugarcane AI. Dive into our Quick Start guide to build your first AI app in no time!
Sugarcane AI is an Open Source Framework : Join our growing community of developers and contributors. Contribute to Sugarcane AI, report issues, or submit feature requests on our GitHub Repository.
Contribute to Sugarcane AI with the following steps:
- Fork this repository.
- Create a new branch for your contributions.
- Implement your changes and commit them with clear messages.
- Push your changes to your forked repository.
- Open a pull request, and we'll review your contributions.
Sugarcane AI's Microservices Architecture is a game-changer for LLM app development, offering developers a streamlined, efficient, and scalable approach.
With a laser focus on Micro LLMs, we provide tools for training, versioning, and shipping these models with ease. You can publish fine-tuned Micro LLMs on Sugar Hub, our collaborative platform, and utilize our Data Assistant to expedite model creation.
This architecture empowers you to harness the full potential of Micro LLMs while simplifying your development journey. It's about clean code, rapid iterations, and improved teamwork, enabling you to supercharge your AI projects.
Microservices foster collaboration among development teams promoting clean and modular code. Separate teams can work on different services concurrently, accelerating development and innovation for top-notch AI apps.
In microservices, prompts, data, and the application layer are decoupled, providing flexibility. Developers can fine-tune LLMs, manage prompts, and handle data independently, enabling rapid experimentation and innovation.
Microservices offer improved reliability, effortless scalability, and streamlined maintenance, ensuring consistent AI app performance without disruptions, regardless of app size, while allowing efficient updates and bug fixes.
Our roadmap for Sugarcane AI is structured around three critical layers within the LLM ecosystem, each designed to empower data scientists and prompt & application developers:
In this phase, we're doubling down on the Prompt Layer. Our mission is to equip you with robust tools for efficient dataset management, prompt control, and Micro LLM training. Expect precise and high-accuracy prompts at your fingertips.
The Marketplace Layer is where the magic happens. We're enhancing this core component to allow you to package, share, and reuse workflows, prompts, and datasets effortlessly across the developer ecosystem, fostering collaboration and innovation across the community.
At the heart of our roadmap lies the LLM Layer, where the magic of Micro LLMs unfolds. We're revamping the way you train and backtest Micro LLMs using labeled datasets, making it more intuitive and efficient. Seamlessly ship your models over APIs, and dive into Sugar Hub, our dedicated platform for sharing and collaborating on fine-tuned Micro LLMs.
- Train and backtest Micro LLMs using labeled datasets
- Version and ship Fine tuned LLMs over APIs
- Publish fine-tuned Micro LLMs on Sugar Hub
Next, we're turning our focus to the Workflow Layer. Brace yourself for a user-friendly UI toolkit that simplifies workflow design, deployment, and integration. We're making it effortless to build workflows into web, mobile, and chat applications.
- Implement workflows/plugins with just a few lines of code
- Client libraries supporting multiple languages
- Integrate managed prompt packages from Sugar Hub
Join us on this exhilarating journey as we transform Sugarcane AI into a powerful and user-friendly platform for Micro LLM-driven workflow automation solutions tailored to developers like you.
Stay updated with the latest news, discussions, and support on our communities.
Have questions, need assistance, personalized support, or just want to share your thoughts on our product? Book an appointment with our team on Calendly here.
Sugarcane AI is licensed under Apache License 2.0 - see the LICENSE file for details.
We appreciate the contributions from our dedicated community members. You make Sugarcane AI shine! ✨
Join Sugarcane AI in simplifying AI app development. Let's make AI accessible to everyone! 🌐 🚀