Local LLM / Local inference setup and APIs
Local LLM: Ollama loaded model inspection
Inspect the running-model endpoint or CLI before diagnosing slow generation.
Environment and inputs: Local inference APIs; Ollama examples require a running local service and a compatible installed model.
When to use it
Make one small request work end to end before adding streaming, tools or batching. Capture the effective model identity and configuration so failures can be reproduced against the server actually used by the application.
Practical example
Read-only calls against an already running local Ollama service. They inspect loaded models; they do not start or download one.
ollama ps
# Equivalent HTTP inventory from the local machine:
curl --fail --silent --show-error http://localhost:11434/api/psCheck the result
Compare reported loaded model, context allocation and GPU placement with the intended configuration.
Save and compare results
Save the test input, configuration and result. Fix the failing boundary, repeat the same test and compare it with the saved working configuration.
Scope and limits
These exercises have not been tested on your systems. Check feature support and configuration names against the installed release and linked documentation.