What is Superduper?
Superduper (formerly SuperDuperDB) is a Python framework designed to integrate AI models and workflows seamlessly with major databases. It enables you to implement custom AI solutions without the hassle of complicated data movements through pipelines and specialized vector databases. Whether you want to host your own models, stream inference, or conduct scalable model training and fine-tuning, Superduper simplifies the entire process.
Transforming your existing database into an AI development and deployment stack is as easy as issuing one command:
db = superduper(mongodb, postgres, mysql, sqlite, duckdb, snowflake:your-db-uri)Imagine a chef in a restaurant kitchen. Instead of having to fetch ingredients from various suppliers across town (like moving your data through complex pipelines), the chef has everything they need already stocked in the kitchen (your existing database). With Superduper, that chef can whip up a variety of creative dishes (custom AI solutions) without ever having to leave the kitchen! It’s about streamlining the cooking process, allowing focus on flavor and presentation instead of logistics.
Key Features
- Integration of AI with your existing data infrastructure: Seamlessly connect AI models with databases without extra processing steps.
- Inference via change-data-capture: Automatically compute outputs as new data comes in, keeping your model updated.
- Scalable model hosting: Host models from HuggingFace, PyTorch, and scikit-learn securely.
- Scalable model training: Train on vast datasets effortlessly and optimize performance.
- Model chaining: Create complex workflows by linking models together.
- Simple Python interface: Reduce boilerplate code with straightforward Python commands.
- Python-first: Utilize any Python function to enhance workflows.
- Difficult data-types: Handle various data formats directly within your database.
- Feature storing: Centralize inputs and outputs of AI models in your database.
- Vector search: Create a multi-modal vector-search database without duplicating data.
Example Use-Cases and Apps
We invite you to explore our notebooks that illustrate various frameworks and techniques. For hands-on ideas on how to use Superduper, check our Docs.
| Name | Link |
|---|---|
| Multimodal vector-search with a range of models and datatypes |  |
| RAG with self-hosted LLM | |
| Fine-tune an LLM on your database | |
| Featurization and transfer learning | |
Installation
Getting started with Superduper is a breeze! Choose your preferred method:
Option 1. Superduper Library
Ideal for building new AI applications.
pip install superduper-frameworkOption 2. Superduper Container
Ideal for learning basic Superduper functionalities and testing notebooks.
docker pull superduperio/superduper
docker run -p 8888:8888 superduperio/superduperOption 3. Superduper Testenv
Ideal for learning advanced functionalities and testing entire AI stacks.
make build_sandbox
make testenv_initTroubleshooting Ideas
If you encounter any problems or have questions:
- Join our Slack community for support.
- Search through our GitHub Discussions for solutions.
- Open a new issue on our GitHub Issues page.
For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.
Conclusion
At fxis.ai, we believe that such advancements are crucial for the future of AI, as they enable more comprehensive and effective solutions. Our team is continually exploring new methodologies to push the envelope in artificial intelligence, ensuring that our clients benefit from the latest technological innovations.