OpenAI-compatible API
The de facto standard shape for talking to a model over HTTP: POST a list of messages to /v1/chat/completions, get a reply back. OpenAI defined it, and every local runner copies it, so LM Studio and Ollama both answer on http://127.0.0.1:1234 or :11434 with the same request format.
This is the single most useful accident in local AI. Any tool written against the OpenAI SDK can be pointed at a local server by changing a base URL, and nothing else has to change.
It also means the choice of model is a runtime decision rather than an architectural one. The same program can run against a 4 GB model on a laptop and a frontier model in the cloud, and swapping between them is one line of config.