Open-source platform for experimenting with LLM pipelines. Build, evaluate, and optimize pipelines to improve accuracy and reduce costs. Includes a toolkit for creating multistep pipelines with customizable parameters. Deployable on your own infrastructure.
Features
Superpipe SDK
Allows you to build multistep pipelines and experiment with parameters like models, prompts, and number of RAG results. It helps in setting up experiments and optimizing for different outcomes.
Superpipe Studio
Offers a graphical user interface to manage and visualize your experiments. You can compare experiments across cost, speed, and accuracy, and observe the details of each pipeline.
Dataset Management
Facilitates building golden sets with easy ground-truth labeling tools, ensuring accuracy and reliability of data used in experiments.
Observability
Enables you to observe pipelines and deep dive into logs for optimization and troubleshooting purposes.