Packages@robojs/ai
PluginOptions
Interface: PluginOptions
Plugin configuration structure resolved during initialization.
Properties
| Property | Type | Description |
|---|---|---|
commands? | boolean | string[] | Command allow/deny list configuration. |
context? | { depth: number; enabled: boolean; } | Surrounding context configuration for understanding ongoing conversations. |
context.depth? | number | Number of recent messages to fetch for context. Default: 8 |
context.enabled? | boolean | Enable surrounding channel context when mentioned. Default: true |
engine? | BaseEngine | Custom AI engine instance to override defaults. |
hooks? | Partial<Record<HookEvent, Hook>> | Hooks to run during engine orchestration events. |
insight? | boolean | Enables vector store insights synchronisation. |
instructions? | string | System instructions injected into AI prompts. |
mcp? | { baseDelayMs: number; extraRetries: number; gracefulDegradation: boolean; maxDelayMs: number; } | MCP error handling configuration. |
mcp.baseDelayMs? | number | Base delay in milliseconds for exponential backoff. Default: 500 |
mcp.extraRetries? | number | Number of extra retry attempts before degrading. Default: 1 |
mcp.gracefulDegradation? | boolean | Enable graceful degradation by removing MCP tools on persistent failures. Default: true |
mcp.maxDelayMs? | number | Maximum delay in milliseconds for exponential backoff. Default: 2000 |
mcpServers? | MCPTool[] | MCP (Model Context Protocol) server configurations for tool integration. |
restrict? | { channelIds: string[]; } | Channel restriction settings limiting where the bot responds. |
restrict.channelIds | string[] | - |
usage? | PluginUsageOptions | Token usage tracking configuration. |
voice? | VoicePluginVoiceOptions | Voice feature configuration delegated to the voice manager. |
whitelist? | { channelIds: string[]; } | Whitelist of channels where the bot can respond freely. |
whitelist.channelIds | string[] | - |