A clean, provider-agnostic bridge over the LLM landscape. Downstream projects talk to a single BridgeEngine and swap models or providers without changing application code. Covers sync, async, and streaming for Ollama, OpenAI, and Anthropic, with a typed conversation buffer, a prompt-template manager, and per-provider exception hierarchy. Built as the shared substrate for my own LLM experiments.
core-llm-bridge
DESTACADO
Capa de abstracción LLM agnóstica del proveedor. Una única interfaz para Ollama, OpenAI y Anthropic — síncrono, asíncrono, streaming — con estado de conversación, tool calling y errores tipados.
Python 3.12+Pydantic v2httpxOllamaOpenAIAnthropic