looplia config

The config command manages your Looplia settings, including user profile and model provider configuration.

Usage

looplia config <subcommand> [options]

Subcommands

Subcommand	Description
`show`	Display current configuration
`topics`	Set topics of interest
`style`	Set writing style preferences
`provider`	Manage model provider settings

User Profile Commands

Show Configuration

looplia config show

Displays your current profile:

{
  "topics": ["AI", "productivity", "technology"],
  "style": {
    "tone": "expert",
    "wordCount": 1500,
    "voice": "first-person"
  }
}

Set Topics

looplia config topics "AI, productivity, writing"

Topics influence content generation and idea synthesis.

Set Style

looplia config style --tone expert --word-count 1500 --voice first-person

Option	Values	Description
`--tone`	`casual`, `professional`, `expert`	Writing tone
`--word-count`	number	Target word count
`--voice`	`first-person`, `third-person`	Narrative voice

Provider Command

Configure which AI provider and model to use for workflow execution.

Show Provider Config

looplia config provider show

Output:

Current Provider Configuration
──────────────────────────────
Provider: anthropic
Model: claude-sonnet-4-20250514
Base URL: https://api.anthropic.com
Auth Token: sk-ant-***...***

Available Presets:
  ANTHROPIC_CLAUDE_HAIKU
  ANTHROPIC_CLAUDE_SONNET
  ZENMUX_MINIMAX_M21
  ZENMUX_GOOGLE_GEMINI3FLASH
  ... (17 total)

Apply a Preset

looplia config provider preset <PRESET_NAME>

Available presets:

Preset	Model	Description
`ANTHROPIC_CLAUDE_HAIKU`	claude-haiku-4-20250514	Fast, cost-effective
`ANTHROPIC_CLAUDE_SONNET`	claude-sonnet-4-20250514	Balanced performance

Preset	Model	Description
`OLLAMA_GLM47_CLOUD`	glm-4.7:cloud	GLM-4.7 cloud model
`OLLAMA_MINIMAX_M21_CLOUD`	minimax-m2.1:cloud	MiniMax-M2.1 cloud

Preset	Model	Description
`OPENROUTER_PRESET`	@preset/looplia-default	User-configured preset

Setup Instructions:

Create preset in OpenRouter dashboard:
- Visit OpenRouter Presets
- Click “Create Preset”
- Name it exactly looplia-default
- Select your preferred model (e.g., Claude Sonnet, GPT-4, Gemini)
- Configure routing options (fallbacks, providers, etc.)
- Save the preset

Configure looplia:

# Set your OpenRouter API key
export OPENROUTER_API_KEY=sk-or-v1-...

# Apply the preset
looplia config provider preset OPENROUTER_PRESET

# Verify configuration
looplia config provider show

Run workflows:

looplia run writing-kit --file article.md

Your workflows will use whatever model you configured in the OpenRouter dashboard. You can change models at any time by updating the preset - no need to reconfigure looplia.

Preset	Model	Provider
`ZENMUX_MINIMAX_M21`	minimax/minimax-m2.1	MiniMax
`ZENMUX_GOOGLE_GEMINI3FLASH`	gemini-3-flash	Google
`ZENMUX_OPENAI_GPT51MINI`	gpt-5.1-codex-mini	OpenAI
`ZENMUX_XAI_GROK41FAST`	grok-4.1-fast	xAI
`ZENMUX_MINIMAX_M21`	minimax-m21	MiniMax
…	…	…

Example:

# Use Claude Haiku for faster, cheaper runs
looplia config provider preset ANTHROPIC_CLAUDE_HAIKU

# Use ZenMux with MiniMax-M2.1
looplia config provider preset ZENMUX_MINIMAX_M21

Set Individual Values

looplia config provider set <key> <value>

Key	Description
`auth-token`	API key for the provider
`model`	Model identifier
`base-url`	API endpoint URL
`provider`	Provider name (`anthropic`, `zenmux`, `openrouter`, `ollama`, `custom`)

Examples:

# Set ZenMux API key
looplia config provider set auth-token sk-zenmux-xxx

# Override model
looplia config provider set model claude-opus-4-20250514

# Custom endpoint
looplia config provider set base-url https://custom-proxy.example.com

Reset Provider Config

looplia config provider reset

Clears all provider settings, reverting to defaults (Anthropic with ANTHROPIC_API_KEY environment variable).

Configuration Files

Settings are stored in ~/.looplia/:

File	Purpose
`user-profile.json`	Topics, style preferences
`looplia.setting.json`	Provider configuration

user-profile.json

{
  "topics": ["AI", "productivity"],
  "style": {
    "tone": "expert",
    "wordCount": 1500,
    "voice": "first-person"
  }
}

looplia.setting.json

{
  "provider": "zenmux",
  "model": "glm-4.7",
  "baseUrl": "https://api.zenmux.ai/v1",
  "authToken": "sk-zenmux-..."
}

Unified Execution Strategy (v0.6.9)

All providers use the same execution pattern for workflow steps:

Component	Description
Skill Executor	Conceptual name for the step execution mechanism
Implementation	Built-in `general-purpose` subagent from Claude Agent SDK
Context Offload	Each workflow step runs in a separate context window

The workflow-executor skill provides the execution protocol that teaches the subagent how to invoke other skills. This unified strategy works across all providers (Anthropic, ZenMux, and custom endpoints).

Environment Variables

Provider settings can also be set via environment variables:

Variable	Description
`ANTHROPIC_API_KEY`	Anthropic API key (default)
`ZENMUX_API_KEY`	ZenMux API key (auto-mapped)
`LOOPLIA_AGENT_MODEL_MAIN`	Override main agent model
`LOOPLIA_AGENT_MODEL_EXECUTOR`	Override executor model

API Key Priority (v0.6.9+): Config file authToken takes precedence over environment variables when explicitly set via looplia config provider set auth-token. Environment variables serve as fallback when no authToken is configured.