Download
Rig is an open source Rust framework designed to help developers build modular and scalable applications powered by large language models. It provides a unified set of abstractions that allow applications to interact with many AI model providers and vector databases through a single interface. Its architecture emphasizes modularity, enabling developers to integrate only the components and integrations they need for a specific application. Rig includes built-in support for agent workflows, allowing systems to perform multi-turn reasoning, tool calling, and retrieval-based tasks within structured pipelines. It also supports capabilities such as text generation, embeddings, transcription, image generation, and audio generation depending on the provider used. Developers can integrate language models into their software with minimal boilerplate while maintaining flexibility for complex AI workflows.
Features
- Unified interface for interacting with more than 20 AI model providers
- Integration support for over 10 vector databases used in semantic search
- Agent workflows capable of multi-turn prompting and tool execution
- Built-in support for completions, embeddings, transcription, and multimodal models
- Streaming responses and observability support using standardized conventions
- Modular architecture with companion crates for providers and vector stores
Project Samples
