mm profile¶

Manage LLM provider profiles — store and switch between API endpoints, models, and keys without editing config files.

Synopsis¶

mm profile SUBCOMMAND [OPTIONS]

Subcommands¶

Subcommand	Description
`list`	List all configured profiles
`add NAME`	Add a new profile
`update NAME`	Update one or more fields of an existing profile
`use NAME`	Switch the active profile
`remove NAME`	Remove a profile

Profile fields¶

Each profile stores three fields:

Field	Description
`base_url`	LLM API base URL (e.g. `https://api.openai.com/v1`)
`model`	Default model name for this endpoint
`api_key`	API key (optional; empty for local providers like Ollama)

Profile resolution order¶

The active profile is resolved in this order of precedence:

--profile NAME flag on the mm command
MM_PROFILE environment variable
active_profile field in ~/.config/mm/profiles.toml
"ollama" (built-in fallback)

mm profile list¶

List all configured profiles with their base_url and model. The active profile is marked with a bullet (●).

mm profile list
mm profile list --format json
mm profile list --format tsv

Options:

Flag	Short	Description
`--format FORMAT`	`-f`	Output format: `rich` (default), `json`, `tsv`, `csv`

API keys are masked (••••) in all output formats.

Example output (rich):

  Profile       base_url                          model
  ──────────────────────────────────────────────────────
● ollama        http://localhost:11434/v1          qwen3-vl:8b
  openai        https://api.openai.com/v1          gpt-4o
  openrouter    https://openrouter.ai/api/v1       Qwen/Qwen3.5-0.8B

mm profile add¶

Add a new named profile. --base-url and --model are required.

mm profile add NAME --base-url URL --model MODEL [--api-key KEY]

Options:

Flag	Short	Description
`--base-url URL`	`-b`	LLM API base URL (required)
`--model MODEL`	`-m`	Default model name (required)
`--api-key KEY`	`-k`	API key (default: empty)

Examples:

# Add an OpenRouter profile
mm profile add openrouter \
  --base-url https://openrouter.ai/api/v1 \
  --model Qwen/Qwen3.5-0.8B

# Add an OpenAI profile with API key
mm profile add openai \
  --base-url https://api.openai.com/v1 \
  --api-key sk-... \
  --model gpt-4o

# Add a local Ollama profile
mm profile add ollama \
  --base-url http://localhost:11434/v1 \
  --model qwen3-vl:8b

# Add a vLLM server
mm profile add vllm \
  --base-url http://gpu-server:8000/v1 \
  --model Qwen/Qwen2.5-VL-7B-Instruct

Returns an error if a profile with the given name already exists.

mm profile update¶

Update one or more fields of an existing profile. Only the fields you specify are changed; others are preserved.

mm profile update NAME [--base-url URL] [--model MODEL] [--api-key KEY]

Options:

Flag	Short	Description
`--base-url URL`	`-b`	New LLM API base URL
`--model MODEL`	`-m`	New model name
`--api-key KEY`	`-k`	New API key

Examples:

# Switch the model on an existing profile
mm profile update ollama --model qwen3.5:0.8

# Rotate an API key
mm profile update openai --api-key sk-new-key

# Update both URL and model
mm profile update openai \
  --base-url https://api.openai.com/v1 \
  --model gpt-4o-mini

Returns an error if the profile does not exist.

mm profile use¶

Switch the active profile. The active profile is persisted in ~/.config/mm/profiles.toml.

mm profile use NAME

Examples:

mm profile use openrouter
mm profile use ollama
mm profile use gemini

Returns an error if the named profile does not exist.

mm profile remove¶

Remove a profile. The currently active profile cannot be removed — switch to another profile first.

mm profile remove NAME

Examples:

mm profile remove openai

Returns an error if the profile is currently active or does not exist.

Per-command profile selection¶

Select a profile for a single command without changing the active profile:

# CLI flag (highest priority)
mm --profile openrouter cat photo.png -m accurate

# Environment variable
MM_PROFILE=openai mm cat photo.png -m accurate

Configuration file¶

Profiles are stored in ~/.config/mm/profiles.toml. The file follows standard TOML format with one section per profile:

active_profile = "ollama"

[ollama]
base_url = "http://localhost:11434/v1"
model = "qwen3-vl:8b"
api_key = ""

[openai]
base_url = "https://api.openai.com/v1"
model = "gpt-4o"
api_key = "sk-..."

Notes¶

Profile names are case-sensitive.
mm profile list --format json emits an {"active": "...", "profiles": {...}} envelope; API keys are masked as "••••" in all formats.
Use mm config reset-profiles to restore all built-in profiles to their defaults and remove custom ones.