A walkthrough of building AI-powered features in Laravel using the Laravel AI SDK, focusing on two key topics: safely handling sensitive business data in LLM prompts by building a filtered 'safe context' layer before sending data to OpenAI or Anthropic, and a new observability dashboard package called Laravel AI Orbit. The safe context approach involves creating service classes that selectively expose only non-sensitive fields to the LLM, combined with system prompt instructions to prevent the model from requesting sensitive data. Laravel AI Orbit provides token usage monitoring, cost tracking, PII detection in prompts, a prompt playground for model comparison, and a reusable prompt library.
•11m watch time
734 Impressions