Project Introduction: Rivet
About Rivet
Rivet is a powerful desktop application designed to create and manage complex AI agents and prompt chaining, which can be seamlessly embedded into other applications. This application supports key artificial intelligence capabilities, including large language models like OpenAI's GPT-3.5 and GPT-4, along with Anthropic's Claude series.
Rivet Application
Rivet is an intuitive desktop application focusing on the construction and deployment of sophisticated AI agents and the management of prompt chaining. The application enables developers to easily integrate these AI solutions into their own applications. Rivet offers support for several prominent AI technologies and services, such as:
- OpenAI: With the inclusion of GPT-3.5 and GPT-4, developers can utilize advanced natural language processing capabilities.
- Anthropic Claude Series: Includes instant Claude, Claude 2, and the more recent Claude 3 with its Haiku, Sonnet, and Opus models.
- AssemblyAI LeMUR: This framework is crucial for applications involving voice data, enhancing the user experience with versatile speech understanding.
To facilitate AI deployments, Rivet also integrates with embedding/vector database services, including OpenAI's Embeddings and Pinecone, ensuring robust support for data-handling within AI models. Moreover, Rivet includes integration for speech-to-text features via AssemblyAI.
Rivet Core
At the heart of Rivet is its core, a robust TypeScript library that allows for the execution and manipulation of graphs built within the Rivet application. Not only is this core utilized by the Rivet application itself, but it also enables developers to incorporate Rivet's functionality into their own software solutions, providing ease of access to Rivet-generated graphs. The Rivet core library is available on NPM under the package names @ironclad/rivet-core
and @ironclad/rivet-node
.
Getting Started
Rivet provides several pathways to get started, catering to different user needs:
- Prebuilt Binaries: Users can download ready-to-use binaries for MacOS, Linux, and Windows, which simplifies the installation process significantly.
- Running from Source: For developers who prefer to run Rivet from its source code, comprehensive steps are provided in the contributing guidelines, which outline the process for building and deploying the application from scratch.
Contributing to Rivet
Contributions are highly valued within the Rivet community, whether they be in the form of code, documentation, bug reports, or feature suggestions. The project follows an inclusive code of conduct to ensure a welcoming environment for all contributors. The community employs the All Contributors bot to acknowledge and celebrate the diverse range of contributions made by individuals.
Troubleshooting and Support
To assist users facing challenges or requiring more information, Rivet provides a support system. Users can visit the GitHub Issues page to find solutions to common problems or to report new issues. Additionally, the Rivet community fosters open discussions regarding general inquiries or innovative ideas related to the project.
Conclusion
With its comprehensive toolset and scalable architecture, Rivet stands as a compelling option for developers looking to integrate sophisticated AI capabilities into their applications. By simplifying the deployment process and providing extensive documentation and support, Rivet is positioned as a user-friendly solution for AI development.