Rivet

Rivet is a visual programming environment for building AI agents using Large Language Models (LLMs). It lets teams construct, debug and collaborate on complex LLM prompt graphs and run them in their own environment.

Rivet helps teams overcome the difficulties of building AI agents programmatically by offering a visual interface that makes it easier to design, debug and collaborate on complex prompt chains. The tool also lets you visualize and build applications for production, not just prototyping. You can debug remotely, seeing prompt chains run in real time, and collaborate with others by versioning Rivet graphs as YAML files in a team's repository.

Some of Rivet's key features include:

• Visualize and Build: Build complex AI agent chains in a visual interface.
• Debug Remotely: See prompt chains run in real time.
• Collaborate: Version and review Rivet graphs in your team's repository.

Rivet is geared for teams building AI-powered apps, where it can help simplify the process of constructing and debugging complex LLM prompt graphs. Rivet can be integrated with Node and TypeScript applications, so it fits in with existing projects.

To get started with Rivet, check the Getting Started guide, which covers building AI agent graphs and integrating Rivet into your application. You can also try the Rivet example application to see how development and debugging works.

Rivet is open-source and runs on MacOS, Linux and Windows. System requirements are MacOS Monterey or later, Windows 10 or later, and a recent version of webkitgtk installed for Linux. It can be built from source with Rust, Node 20+, and yarn.

For more information on Rivet, including its API reference and user guide, check the Rivet website.

Published on June 13, 2024