Sample Configuration File
In this page you find a full configuration with all sections commented out. All sections and keys are optional, uncomment what fits your setting.
toml
#"############### promptcmd ################
#
# To modify, uncomment any section to edit
# it configuration.
#
##########################################
### Configuration for the "create" command
##########################################
# [create]
# no_enable = false # Auto enable prompt once created
# force = false # Force save prompt files disregarding validation result
##########################################
### Configuration for the "import" command
##########################################
# [import]
# enable = false # Auto enable prompt once imported
# force = false # Force import disregarding validation result
###########################################
### Default Configuration for all providers
###########################################
# [providers]
### Default model to use if not specified in prompt files.
### Can be in provider/model format, just the provider name,
### a variant, or a group name.
# default = "ollama/gpt-oss:20b"
# temperature = 1.0
# system = "You are a useful assistant"
# max_tokens = 1000
# cache_ttl = 0 # Number of seconds to cache responses
#################################
### GenAI Providers Configuration
#################################
# [providers.openai]
# model = "gpt-5-mini-2025-08-07"
# api_key = "sk-proj-xxxx"
# temperature = 1.0
# max_tokens = 1.0
# system = "You are a useful assistant"
# cache_ttl = 0
# [providers.openrouter]
# model = "anthropic/claude-sonnet-4"
# api_key = "sk-or-xxxx"
# temperature = 1.0
# max_tokens = 1.0
# system = "You are a useful assistant"
# cache_ttl = 0
# [providers.google]
# model = "gemini-2.5-flash"
# api_key = "aaaaaa..."
# temperature = 1.0
# max_tokens = 1.0
# system = "You are a useful assistant"
# cache_ttl = 0
# [providers.anthropic]
# api_key = "sk-ant-xxxx"
# model = "claude-opus-4-5"
# temperature = 1.0
# max_tokens = 1.0
# system = "You are a useful assistant"
# cache_ttl = 0
# [providers.ollama]
# endpoint = "http://127.0.0.1:11434"
# model = "gpt-oss:20b"
# cache_ttl = 0
#########################################################################
### Configurations for Variants.
### These inherit their provider's configuration, overriding any property
### as needed
#########################################################################
# [providers.anthropic.rust-coder]
# system = """You are a rust coding assistant helping me with rust questions.
# Be brief, do not use markdown in your answers. Prefer to answer with pure code
# (no before and after explanation unless very appropriate)"""
##########################################################################
### Configurations for Groups.
### Executions are load balanced across members of a group based on
### token consumption. Load is split proportionate to any indicated weight
###########################################################################
# [groups.balanced]
# providers = [
# "anthropic", "google", "openrouter"
# ]
# [groups.unbalanced]
# providers = [
# { name = "google", weight = 5 },
# { name = "anthropic", weight = 1 },
# ]