OpenClaw

Learn about configurations >

• ai-api-key | string

  API key for the selected AI provider. Required for most providers:

  • anthropic: Anthropic API key
  • openai: OpenAI API key
  • google: Google Gemini API key
  • opencode: OpenCode API key
  • github-copilot: GitHub personal access token
  • openrouter: OpenRouter API key
  • xai: xAI API key
  • groq: Groq API key
  • cerebras: Cerebras API key
  • mistral: Mistral API key
  • zai: Z.AI API key
  • vercel-ai-gateway: Vercel AI Gateway API key Not required for: ollama (local), bedrock (AWS credentials), openai-codex (OAuth)

  Note: API keys are stored securely in auth-profiles.json, not in environment variables.

• ai-base-url | string

  Custom API base URL for OpenAI-compatible providers. Use this to connect to local AI services like LM Studio, vLLM, FastChat, etc. Examples:

  • LM Studio: http://localhost:1234/v1
  • vLLM: http://localhost:8000/v1
  • Text Generation WebUI: http://localhost:5000/v1 Leave empty to use default provider endpoints.

• ai-model | string

  AI model(s) to use. Supports:

  • Single model: claude-opus-4-5
  • Multiple models (comma-separated): gemini-3-flash-preview,gemini-3-pro-preview,gemini-2.5-pro

  When multiple models are provided:

  • First model becomes the primary (system default)
  • Remaining models are added as fallbacks
  • All models use the configured ai-provider

  Models are stored as provider/model format (e.g., anthropic/claude-opus-4-5).

  For models from different providers, use separate slots (ai0-provider/ai0-model, etc.) with each provider's API key.

  Examples:

  • Single: claude-opus-4-5
  • Multiple same provider: gpt-4,gpt-3.5-turbo,gpt-4o-mini
  • GitHub Copilot aggregator: gemini-3-flash-preview,claude-haiku-4.5,gpt-4

• ai-provider | string

  Primary AI model provider. Supported providers:

  • anthropic: Anthropic Claude models
  • openai: OpenAI GPT models
  • openai-codex: OpenAI Codex (requires OAuth)
  • google: Google Gemini models
  • opencode: OpenCode Zen models
  • github-copilot: GitHub Copilot/Models API
  • openrouter: OpenRouter aggregator
  • xai: xAI models
  • groq: Groq models
  • cerebras: Cerebras models
  • mistral: Mistral AI models
  • zai: Z.AI (GLM) models
  • vercel-ai-gateway: Vercel AI Gateway
  • ollama: Local Ollama models (no API key needed)
  • bedrock: AWS Bedrock (uses AWS credentials)

• ai0-api-key | string

  API key for AI model slot 0

• ai0-base-url | string

  Custom API base URL for AI model slot 0 (for OpenAI-compatible local services)

• ai0-model | string

  AI model(s) for slot 0. Supports single or comma-separated models. All models in this slot use ai0-provider. Example: gpt-4,gpt-3.5-turbo,gpt-4o-mini

• ai0-provider | string

  AI provider for model slot 0. See ai-provider for supported providers.

• ai1-api-key | string

  API key for AI model slot 1

• ai1-base-url | string

  Custom API base URL for AI model slot 1 (for OpenAI-compatible local services)

• ai1-model | string

  AI model(s) for slot 1 (supports comma-separated)

• ai1-provider | string

  AI provider for model slot 1. See ai-provider for supported providers.

• ai2-api-key | string

  API key for AI model slot 2

• ai2-base-url | string

  Custom API base URL for AI model slot 2 (for OpenAI-compatible local services)

• ai2-model | string

  AI model(s) for slot 2 (supports comma-separated)

• ai2-provider | string

  AI provider for model slot 2. See ai-provider for supported providers.

• ai3-api-key | string

  API key for AI model slot 3

• ai3-base-url | string

  Custom API base URL for AI model slot 3 (for OpenAI-compatible local services)

• ai3-model | string

  AI model(s) for slot 3 (supports comma-separated)

• ai3-provider | string

  AI provider for model slot 3. See ai-provider for supported providers.

• ai4-api-key | string

  API key for AI model slot 4

• ai4-base-url | string

  Custom API base URL for AI model slot 4 (for OpenAI-compatible local services)

• ai4-model | string

  AI model(s) for slot 4 (supports comma-separated)

• ai4-provider | string

  AI provider for model slot 4. See ai-provider for supported providers.

• ai5-api-key | string

  API key for AI model slot 5

• ai5-base-url | string

  Custom API base URL for AI model slot 5 (for OpenAI-compatible local services)

• ai5-model | string

  AI model(s) for slot 5 (supports comma-separated)

• ai5-provider | string

  AI provider for model slot 5. See ai-provider for supported providers.

• ai6-api-key | string

  API key for AI model slot 6

• ai6-base-url | string

  Custom API base URL for AI model slot 6 (for OpenAI-compatible local services)

• ai6-model | string

  AI model(s) for slot 6 (supports comma-separated)

• ai6-provider | string

  AI provider for model slot 6. See ai-provider for supported providers.

• ai7-api-key | string

  API key for AI model slot 7

• ai7-base-url | string

  Custom API base URL for AI model slot 7 (for OpenAI-compatible local services)

• ai7-model | string

  AI model(s) for slot 7 (supports comma-separated)

• ai7-provider | string

  AI provider for model slot 7. See ai-provider for supported providers.

• ai8-api-key | string

  API key for AI model slot 8

• ai8-base-url | string

  Custom API base URL for AI model slot 8 (for OpenAI-compatible local services)

• ai8-model | string

  AI model(s) for slot 8 (supports comma-separated)

• ai8-provider | string

  AI provider for model slot 8. See ai-provider for supported providers.

• ai9-api-key | string

  API key for AI model slot 9

• ai9-base-url | string

  Custom API base URL for AI model slot 9 (for OpenAI-compatible local services)

• ai9-model | string

  AI model(s) for slot 9 (supports comma-separated)

• ai9-provider | string

  AI provider for model slot 9. See ai-provider for supported providers.

• auto-update | boolean

  Automatically update OpenClaw to latest stable version on upgrade-charm

• discord-bot-token | string

  Discord bot token from Discord Developer Portal. Leave empty to disable Discord messaging.

• dm-policy | string

  Default: pairing

  DM access policy: 'pairing' (require pairing code), 'open' (auto-respond), or 'closed' (reject all DMs)

• dm-scope | string

  Default: main

  DM session scope for isolating direct message conversations:

  • 'main' (default): All DMs share the main session for continuity across devices/channels
  • 'per-peer': Isolate sessions by sender ID across channels
  • 'per-channel-peer': Isolate by channel + sender (recommended for multi-user inboxes)
  • 'per-account-channel-peer': Isolate by account + channel + sender (recommended for multi-account inboxes)

  Security Warning: If your agent receives DMs from multiple people, use 'per-channel-peer' or 'per-account-channel-peer' to prevent conversation context leakage between users.

• gateway-bind | string

  Default: loopback

  Gateway bind mode: 'loopback' (127.0.0.1 only), 'lan' (all interfaces), or specific IP address.

  IMPORTANT: Multi-unit deployments REQUIRE 'lan' mode for Node connectivity. The charm will block deployment if loopback mode is used with multiple units.

• gateway-port | int

  Default: 18789

  Port for the OpenClaw Gateway WebSocket/HTTP server

• install-method | string

  Default: npm

  Installation method: 'npm' (global npm install), 'pnpm' (uses npm), 'bun' (global bun install), or 'source' (build from git source)

• line-channel-access-token | string

  LINE Messaging API channel access token from LINE Developers Console. Leave empty to disable LINE messaging.

• line-channel-secret | string

  LINE Messaging API channel secret from LINE Developers Console. Required when line-channel-access-token is configured.

• log-level | string

  Default: info

  Log level (debug, info, warn, error)

• manual | boolean

  Manual configuration mode. When enabled (manual=true):

  • Charm will NOT auto-generate openclaw.json or node.json
  • User is responsible for creating and managing OpenClaw configuration
  • Charm only installs OpenClaw and manages systemd services
  • Peer relations still work for multi-unit deployments
  • Environment file (.openclaw/environment) is still created for systemd

  When disabled (manual=false, default):

  • Charm auto-generates configuration from Juju config options
  • AI providers, channels, and settings are managed by the charm
  • Configuration updates on 'juju config' changes

• node-version | string

  Default: 24

  Node.js major version to install (minimum 22 required)

• sandbox-mode | string

  Default: non-main

  Sandbox mode for non-main sessions: 'all' (sandbox everything), 'non-main' (sandbox groups/channels), or 'none' (no sandboxing)

• slack-app-token | string

  Slack app token (xapp-...) from Slack App settings. Required when slack-bot-token is configured.

• slack-bot-token | string

  Slack bot token (xoxb-...) from Slack App settings. Leave empty to disable Slack messaging.

• telegram-bot-token | string

  Telegram bot token from @BotFather (e.g., 123456:ABC-DEF1234ghIkl-zyx57W2v1u123ew11). Leave empty to disable Telegram messaging.

• use-browser | string

  Browser to install for browser automation. OpenClaw's browser tool enables web browsing, screenshots, and automation via Playwright. Supported values: 'chrome', 'chromium', 'firefox', or empty string to disable. Can be changed post-deployment via 'juju config'.

• version | string

  Default: latest

  OpenClaw version to install (e.g., 'latest', '2026.1.29', or git commit)