dreadnode/rigging
Lightweight LLM Interaction Framework
What it solves
Rigging simplifies the integration of Large Language Models (LLMs) into production code. It removes the boilerplate associated with managing different model providers, handling structured data, and building complex chat pipelines, making the process of using AI in software more efficient and predictable.
How it works
Rigging acts as a lightweight framework that leverages LiteLLM, vLLM, and transformers to provide a unified interface for a wide array of models. It allows developers to define prompts as Python functions with type hints and docstrings, and uses Pydantic models for structured output validation. It manages model configurations via connection strings (similar to database connections) and supports async batching for high-scale generation.
Who it’s for
Developers building AI agents or LLM-powered applications who need a flexible, type-safe, and production-ready library to manage prompts, tool use, and model switching.
Highlights
- Unified Model Access: Instant access to numerous models via LiteLLM, vLLM, and transformers.
- Structured Data: Seamlessly switch between unstructured text and Pydantic models for validated output.
- Prompt Engineering: Define prompts as Python functions with built-in type hints.
- Production Tooling: Integrated tracing with Logfire, async batching, and serialization support.
- Flexible Tool Use: Simple tool integration, even for models that do not natively support tool-calling APIs.