What problem does it solve? Setting up Gemini AI in a Replit project normally requires obtaining and managing your own API key. This Skill provisions Gemini access through the Replit AI Integrations proxy with automatically configured environment variables, so you can add chat, streaming, and image generation features without handling credentials. ## Core Features & Use Cases - Managed Gemini Access: Provisions AI_INTEGRATIONS_GEMINI_BASE_URL and AI_INTEGRATIONS_GEMINI_API_KEY automatically, billing usage to Replit credits. - Chat with SSE Streaming: Provides a pre-configured GoogleGenAI client, Drizzle schemas for conversations and messages tables, and OpenAPI contract entries for streaming chat endpoints. - Image Generation: Includes a generateImage helper using gemini-2.5-flash-image that returns base64 image data. - Batch Processing: Ships batchProcess and batchProcessWithSSE utilities with rate limiting, retries, and SSE progress events for high-volume LLM tasks. - Use Case: Build a full-stack chat application where users create conversations, send messages, and receive streamed Gemini responses persisted in a PostgreSQL database via Drizzle ORM. ## Quick Start Ask the agent to set up the Gemini AI integration and build a chat endpoint that streams responses using the provided client and batch utilities.
