Nouli – A decision model TUI to branch and blend multiple histories (github.com)

Nouli has launched a new decision model tool that enables users to track and manage decision histories seamlessly within a terminal interface. This application, structured as a CLI and Rust library, allows for an interactive ledger-backed decision network, where users can add facts, frame questions, and execute model requests without losing track of previous decisions. It supports multiple decision service integrations and features five task-specific models available through Hugging Face, including classifiers for banking intents and hate speech detection. This tool is significant for the AI/ML community as it emphasizes not just decision-making but also the contextual history behind those decisions, allowing for iterative improvements and exploration of alternatives. Nouli's architecture promotes a transparent decision-making process by making it possible to branch and blend decision paths while preserving original contexts. With features like reusable saved answers and local model serving recipes, Nouli enhances efficiency for developers and data scientists tackling complex decision-based tasks in various sectors, making it a powerful addition to the toolkit for AI-driven applications.