Revolutionizing Computational Notebooks with Riverbox

Riverbox is a new flow-based notebook designed to address key pain points of traditional computational notebooks, making them deployable, collaborative, and portable.

Addressing Key Pain Points of Traditional Notebooks

Deployability

Traditional notebooks lack a standard execution flow and request/response interface. Riverbox introduces input and output cells, making notebooks natively deployable with JSON inputs and outputs.

Collaboration & Code Reuse

No more copy-pasting! Riverbox enables modularization and allows you to "call" an external notebook as a cell within another, solving code sharing issues.

Portability

Riverbox simplifies packaging notebooks and ensures they run in the intended environment, including packages, filesystem, OS, and machine.

Native Parallelization

With flow-based connections and "for each" edges, Riverbox natively supports running cells in parallel for multiple inputs or different code cells simultaneously.

Resource Efficiency (Serverless)

Leveraging a serverless implementation, Riverbox offers a dynamic pool of resources and a pay-as-you-use model, reducing resource waste.

How Riverbox Transforms Your Workflow

Riverbox changes the way you interact with computational notebooks, moving towards a more efficient, scalable, and collaborative environment.

Flow-based Design

Cells connect via edges, including "for each" edges, enabling intuitive data flow and native parallelization.

Modular and Reusable

Define input and output cells for your notebooks, making them callable and reusable components, fostering better code organization.

Serverless Execution

Benefit from a dynamic resource pool and efficient scheduling, leading to cost-effective execution on demand.

Environment Consistency

When calling external notebooks, Riverbox ensures the original author's environment is used, guaranteeing consistent results.

Flexible Resource Allocation

Specify resource demands (e.g., GPU) for your notebooks, and Riverbox's orchestrator handles the allocation automatically.

HTTP API Integration

Every Riverbox can be invoked via an HTTP API, providing seamless integration with other applications and workflows.

Experience the Future of Computational Notebooks

Riverbox is designed to solve the real-world challenges faced by software engineers and data scientists using computational notebooks today. We've reimagined the notebook format to be inherently more powerful and versatile.

Our focus on deployability, collaboration, and portability, combined with a serverless architecture, provides a robust and efficient platform for all your computational needs.

Unlock the Full Potential of Your Code

With Riverbox, you can easily create complex computational flows, leverage parallel execution, and share your work effortlessly. Imagine running multiple inference calls for various inputs simultaneously, all managed seamlessly by Riverbox's orchestrator.

Our system design ensures efficient resource utilization, offering a cost-effective solution for diverse computational demands, from simple scripts to large-scale machine learning models.

Learn more about Riverbox capabilities