Generate AI-powered welcome messages using Genkit and LLMs. From simple prompts to smart moderation—all streaming in real-time via SSE.
Watch how each flow version handles different use cases, from simple text generation to advanced AI-moderated content with natural language understanding.
Basic string input evolving to rich structured parameters with language, tone, and length control.
Simple to Structured Input
AI-powered text generation
Returns rich metadata alongside generated content for complete transparency and debugging.
Metadata Output
Rich structured responses
Multi-step flow with content safety checking, toxicity filtering, and automatic sanitization.
Content Moderation
AI-powered safety filtering
AI interprets free-form descriptions, extracts parameters, generates content, and moderates—all in one flow.
Smart Interpretation
Natural language understanding
Explore our progressive implementations from simple string I/O to advanced AI-moderated smart flows. Each version builds on the previous, showcasing production-ready patterns.
Enter any occasion or context, and we'll generate a welcome note. This version is intentionally simple and sends your text directly to the AI.
Demo only. Text you enter is sent directly to the AI model and may produce unexpected or nonsensical output, especially for unusual or nonsensical inputs.
Provide a specific occasion and customize the welcome note with language, length, and tone. This version uses structured inputs to give you more control.
Demo only. Your text and selections are sent directly to the AI model. Unusual or unclear inputs may still produce creative or unexpected results.
Same as V2, but the flow returns a structured JSON response: the welcome note plus metadata about how it was generated (interpreted occasion, tone, sentiment, safety, etc.).
Demo only. Your text and selections are sent directly to the AI model. The response is parsed into typed JSON on the backend so you can inspect both the note and its metadata.
Includes automatic content safety checking and sanitization. Try requesting toxic or inappropriate content to see moderation in action.
Original text before sanitization:
No safe version could be generated from the original text.
This note passed content safety filters. No changes were required.
Just describe what you want in plain English. The AI will interpret your request, generate the note, and moderate it.
Original text before sanitization:
No safe version could be generated from the original text.
This note passed content safety filters. No changes were required.