Player not loading? Watch on YouTube
Uploaded on September 9, 2023, this tutorial introduces Rivet as an open source visual editor for building LLM prompt graphs. The source gives no Rivet version or separate recording date. Its installation requirements and interface details describe that period.
The presenter compares Rivet with Langflow and Flowise, noting their similar drag-and-drop interfaces. He argues that Rivet's remote debugging, embedding in a host application, and TypeScript compatibility give it more flexibility. This is a feature discussion rather than a measured comparison.
The walkthrough installs Rivet on Windows using a downloadable installer. For building from source, the presenter lists Rust, Node 20+, Yarn, and Git as prerequisites. He then explains visual prompt chains, observing execution through the remote debugger, and storing graphs as YAML files for team version control.
The example setup requires an OpenAI API key. A chemistry workflow uses GPT-4 to process file context and produce JSON, while the editor demonstration selects GPT-3.5 Turbo and discusses temperature settings and system prompts. Installing the editor locally does not establish offline model inference: the demonstrated workflow relies on OpenAI API access.
The final section introduces graph creation, connecting nodes, JSON export, and plugin options. The demonstration covers basic configuration rather than a complete, tested complex AI agent.
