Which models can Bridge run?

Any model your private runtime serves — none of them in the cloud.

Updated August 2026

Bridge is the harness, not the model

Alluqi Bridge connects to model runtimes you control — an Ollama server, a vLLM endpoint, or another private inference endpoint on your machine or your network. Whatever models that runtime serves, Bridge can use.

No cloud models, by design

Bridge does not send your work to cloud AI providers. That is the product’s core promise: prompts, files, and outputs stay on hardware you control. The one nuance is hybrid web search, where your local model can fetch fresh information from the web — explained in Hybrid web search.

Choosing a model

Different work wants different models — small fast ones for quick questions, larger ones for reasoning over documents or code. The Learn site’s Models + runtimes guide covers the tradeoffs.