Class AIError

constructor

• new AIError(options: AIErrorOptions): AIError

  Creates a new AIError instance.

  Parameters

  • options: AIErrorOptions

    AI error construction options including provider context.

  Returns AIError

  Example

  import { AIError } from '#core/errors/ai-error.js';
  
  const error = new AIError({
    message: 'Token limit exceeded for GPT-4o',
    attempted: 'Generate test steps from intent',
    provider: 'azure-openai',
    model: 'gpt-4o',
    tokenUsage: { prompt: 12000, completion: 0, total: 12000 },
  });
  Copy

Readonlymodel

Readonlyprovider

ReadonlytokenUsage

toAIContext

• toAIContext(): AIErrorContext & {
      model: string
      | undefined;
      provider: string | undefined;
      tokenUsage: TokenUsage | undefined;
  }

  Returns structured context for AI agents with provider diagnostics.

  Returns AIErrorContext & {
      model: string | undefined;
      provider: string | undefined;
      tokenUsage: TokenUsage | undefined;
  }

  Base AI context plus provider, model, and tokenUsage fields to enable LLM self-healing and retry decisions.

  Example

  const context = error.toAIContext();
  // context.provider, context.model, context.tokenUsage available
  // Send to LLM for provider fallback or token budget adjustment
  Copy

toJSON

• toJSON(): SerializedPramanError & {
      model: string
      | undefined;
      provider: string | undefined;
      tokenUsage: TokenUsage | undefined;
  }

  Serializes the error to a JSON-safe object with AI provider fields.

  Returns SerializedPramanError & {
      model: string | undefined;
      provider: string | undefined;
      tokenUsage: TokenUsage | undefined;
  }

  Base fields plus provider, model, and tokenUsage.

  Example

  const json = error.toJSON();
  // json.provider === 'azure-openai'
  // json.model === 'gpt-4o'
  // json.tokenUsage === { prompt: 12000, completion: 0, total: 12000 }
  Copy