Rivet
Product · Developer tool by Ironclad
The Open-Source Visual AI Programming Environment
An open-source visual programming environment for building and debugging complex AI agents with LLMs. It enables developers to design prompt graphs visually and run them directly in Node.js or TypeScript applications.
- Price
- From 0 USD
- Languages
- TypeScript, Node.js
Category: Developer Tools
- Programming languages
- Node.js, TypeScript
- Integrations
- AssemblyAI, Node.js, TypeScript
- License
- Open Source
- Deployment
- Self-hosted, integrated into local applications
- Primary Interface
- Visual, node-based graph editor
- Supported Environments
- Node.js, TypeScript
- Open-source visual AI programming environment
- Node-based editor for building complex LLM prompt graphs
- Real-time remote debugger for observing execution in applications
- Version-controlled graphs stored as YAML files
- Native support for Node.js and TypeScript integrations
Overview
Rivet is an open-source visual programming environment developed by Ironclad for building AI agents with Large Language Models (LLMs). It provides a node-based graph editor that allows developers to design and iterate on complex prompt chains visually rather than programmatically. The tool features a remote debugger for real-time observation of prompt execution within applications, helping teams identify and fix issues in agentic logic. Rivet stores graphs as YAML files, enabling standard version control and collaborative code reviews. It is designed to be integrated directly into Node.js or TypeScript environments, allowing developers to deploy AI workflows in their own production infrastructure. Ironclad uses Rivet internally to power its own Contract AI capabilities.
Visual AI Programming
- Visual Graph Editor: Developers can design and iterate on complex LLM prompt graphs using a visual, node-based programming interface.
- Production-Ready Chains: The tool allows for the creation of complex chains suitable for production applications, moving beyond simple prototyping.
- Collaborative Workflows: Prompt graphs are stored as YAML files, allowing teams to version them in repositories and use standard code review tools.
- AI Agent Construction: The environment supports building AI agents that can be integrated directly into applications.
Debugging and Execution
- Real-Time Observation: Developers can observe the execution of prompt chains within their own application in real-time to identify issues.
- Execution Visibility: The debugger provides visibility into what is happening "under the hood" during the execution of agentic logic.
- Remote Execution: The remote executor allows prompt graphs to run directly in the developer's own environment.
Integration and Deployment
- Node.js and TypeScript Support: Rivet can be integrated into Node.js or TypeScript applications to run the designed prompt graphs.
- API and Tool Integration: The tool is designed to combine AI with existing tools and APIs for robust application building.
- Example Application Integration: Developers can use the Rivet example application to understand how to develop and debug chat applications.
Use cases
- Developers build complex LLM prompt graphs visually to create production-ready AI applications.
- Engineering teams version-control AI logic by storing Rivet graphs as YAML files in their standard repositories.
- Software engineers debug agentic logic in real-time by observing prompt chain execution within their own applications.
- Teams integrate AI-powered experiences like audio transcription and chat into their existing Node.js or TypeScript software.
Specifications
- Programming languages
- Node.js, TypeScript
- Integrations
- AssemblyAI, Node.js, TypeScript
- License
- Open Source
- Deployment
- Self-hosted, integrated into local applications
- Primary Interface
- Visual, node-based graph editor
- Supported Environments
- Node.js, TypeScript
- Storage Format
- YAML
- Debugging Mode
- Real-time remote debugging
- Developer Tools
- Remote executor, example chat application
How it works
- Build Agent Graphs: Developers follow the Getting Started guide to learn how to construct AI agent graphs using the visual editor.
- Integrate into Applications: The designed prompt graphs are integrated into Node.js or TypeScript applications using the remote executor.
- Debug and Iterate: Developers use the remote debugger to observe real-time execution and iterate on the logic within their own environment.
Pricing
- Open Source
- Free
FAQ
What is Rivet?
Rivet is an open-source visual programming environment specifically designed for building and debugging complex AI agents using Large Language Models (LLMs).
Can I debug my AI agents in real-time?
Rivet provides a remote debugger that allows developers to see what is happening "under the hood" and observe the execution of prompt chains in their application in real-time.
Can I run Rivet graphs in my own application?
Yes, Rivet is designed to be integrated into Node.js or TypeScript applications, allowing you to run your prompt graphs directly in your own environment.
Added by AntSpace from public sources (rivet.ironcladapp.com).