unchained-ai

To build LLM pipelines, you use unchained-ai's pipeline primitives to define sequential execution steps. You implement these primitive traits within the monorepo to create scalable, auditable end-to-end LLM workflows.

A provider registry catalogs available LLM providers and their models. You add new models by extending provider interfaces and running the model generator to automatically generate the required model metadata for integration.

Monitoring agent status involves querying the agentic platform limits and using unchained-ai's built-in status monitoring tooling. This ensures your LLM workflow maintains operational visibility throughout pipeline execution.

Yes, rig-core tooling integrates with unchained-ai LLM pipelines to support core interfaces and usage patterns. This integration allows you to extend pipeline components and generate model metadata within your workflow architecture.

You run the model generator when adding new providers or models to the provider registry. It generates the necessary model metadata to ensure proper pipeline integration and accurate model cataloging across the monorepo.