GemChat

GemChat is a collection of small AI experiments (CLI chats, LangGraph agents, and image/story generators). Everything runs locally via simple scripts.

Disclaimer: This project is not affiliated with, endorsed by, or associated with Google or the Gemini AI team. It is an independent project created for development and testing purposes only.

Prerequisites

• Node.js v18 or higher
• npm v8 or higher
• A Gemini API key (Get one for free at https://makersuite.google.com/app/apikey)

Project layout (one-level folders)

• cli-chats/
  • gemini-streaming.ts: streaming Gemini CLI chat (npm run chat:gemini:stream)
  • gemini-with-context.ts: Gemini + LangChain with fixed context (npm run chat:gemini:context)
  • gemini-with-history.ts: Gemini + LangChain chat history (npm run chat:gemini:history)
  • azure-openai.ts: Azure OpenAI CLI chat (npm run chat:azure)
• agents/
  • blog-langraph-basic.ts: simple end-to-end blog writer with Together cover image
  • blog-langraph-with-hitl.ts: multi-step approval loop for outline/title/content/cover
• image-generation/
  • gemini-image.js: Gemini native image generation sample
  • together-image.js: Together FLUX image generation sample
• storybooks/
  • gemini-storybook.js: Gemini storybook generator + per-page images

Setup

1. Install deps: npm install
2. Create .env with any keys you need:
   • GEMINI_API_KEY (Gemini + LangChain samples)
   • TOGETHER_API_KEY (Together + LangGraph cover/image flows)
   • AZURE_OPENAI_API_KEY, AZURE_OPENAI_API_VERSION, AZURE_OPENAI_RESOURCE_NAME (Azure chat sample)

Running scripts

• Gemini streaming chat: npm run chat:gemini:stream
• Gemini LangChain context chat: npm run chat:gemini:context
• Gemini LangChain history chat: npm run chat:gemini:history
• Azure OpenAI chat: npm run chat:azure
• Blog agents: npm run agent:blog:basic or npm run agent:blog:advanced
• Gemini storybook: npm run story:gemini
• Image generation: npm run image:gemini or npm run image:together