provider-adapter

To create a new LLM provider adapter in NestJS, scaffold the adapter class with core methods including formatRequest, buildHeaders, buildUrl, parseResponse, and fetchModels, then register it in the AdapterFactory to enable standardized API formatting.

The BaseProviderAdapter pattern is a standardized structure for LLM integration that defines required methods including formatRequest, buildHeaders, buildUrl, parseResponse, and fetchModels to ensure consistent interfaces across different providers.

To add support for OpenAI-compatible providers like Groq or Together AI, generate an adapter skeleton using the scaffold, wire it into the backend module, and register it in the AdapterFactory to enable immediate model discovery and standardized API formatting.

Yes, you can use TypeScript to build custom LLM provider adapters. The scaffold generates adapter skeletons in TypeScript, implementing core methods and wiring steps required for production-ready adapters with standardized API formatting.

To register a new adapter in the AdapterFactory, add the provider to the ProviderType enum, implement the adapter class with core methods, register it in the factory, and expose it through the backend module for immediate use.

Yes, scaffolding LLM adapters works for mock providers. The scaffold supports OpenAI-compatible, custom, and mock providers, applying consistent interfaces and core methods for production-ready adapter creation.