Ollama, vLLM and llama.cpp all speak OpenAI-style tool calling, but each has a default that can wreck an agent loop, from a 4k context window to unconstrained arguments. Here is what to change before you blame the model.
Running an agent on a local model: the tool-calling settings that quietly break it
calendar_today
November 2, 2026
domain
agenttoolbox-net