Inside Isaac Juracich’s AI tools: Path and Anvil

AI engineering includes the systems that run a request and the interface where a person decides what to do with its result. Two Juracich projects, Path and Anvil, address those different layers. Together they illustrate the range of my work on developer tools.

Path: connecting an application to AI execution

Path provides an API layer for applications that need to connect to supported AI backends. Its public product page describes its compatibility and setup. The engineering problem is giving the calling application a consistent way to request work and receive its result.

For a product using AI, that boundary matters. The interface should not have to reproduce provider-specific execution details in every feature. It needs a way to understand the request’s state and handle an unavailable service. The Path overview is the place to check the current supported setup before choosing it for a project.

Anvil: assistance inside a real terminal

Anvil is a desktop terminal built with Electron and xterm. Its public design centers on asking questions about terminal context and receiving commands that the user can choose to run. The assistant does not independently execute those suggestions.

The product also masks credentials in context before sending it to a model. That is a specific protective measure, not a reason to assume any arbitrary terminal session is safe to share. The boundary between local context and a remote service remains an important part of using an assistant.

Control is part of the interface

A suggestion and an executed action are different events. Anvil makes that distinction visible by leaving execution to the person at the keyboard. This provides a concrete example of how an AI feature can assist an existing workflow while preserving a clear decision point.

The broader design question for any assistant is similar: what can it read, what may it propose, and which actions need the user’s involvement? Those choices should be made before connecting a model to business tools.

What this work demonstrates

These projects connect API design, desktop interfaces, model integration, and product behavior. They sit alongside my mobile application experience and business automation work.

Visit the product pages for current capabilities, or discuss a focused AI feature for your own system.