NAME

INSPECT — show detailed Ollama model information

SYNOPSIS

/inspect /inspect MODEL

DESCRIPTION

/inspect queries the local Ollama server for detailed information about installed models. Requires an Ollama backend; use /model use to pick an Ollama model, and run “ollama serve” in a shell if Ollama is not running.

Without a MODEL argument, /inspect shows a summary table of all installed models: name, disk size, family, context length, and capability flags (tools, embed, tagged-blocks).

With a MODEL argument, /inspect shows the full detail view for that model: family, parameter count, quantization level, disk size, context length, and all Modelfile parameter lines (e.g. temperature, system prompt).

EXAMPLES

Summary of all installed models:

  harvey > /inspect

Detail view for a specific model:

  harvey > /inspect gemma4:e2b
  Model:        gemma4:e2b [loaded]
  Family:       gemma4
  Parameters:   2.5B
  Quantization: Q4_K_M
  Context:      131072 tokens
  Disk size:    1.7 GiB

  Modelfile parameters:
    stop "<end_of_turn>"

SEE ALSO

/model list — models across llamafile, llama.cpp, and Ollama /model use NAME — switch models (an Ollama model is probed on first alias) ollama show MODEL — raw Modelfile, run in a shell /help model