The LLM Claude cannot generate images by itself, but it can tap on external tools to do so. Today, I tried out the Qwen and Krea image engines to make some fun images that included both English and Chinese text.
While I usually recommend people to use ChatGPT and Google Whisk to generate visuals, the open-source models like Qwen are advancing rapidly to improve text generation and image realism. I need to find some time to download Qwen to my computer and see how well it works with local AI processing.
Tech talk: Of the three major LLMs, Claude offers the greatest flexibility. Its maker Anthropic pioneered the Model Context Protocol (MCP) to connect LLMs with different external tools. The tools (such as image engines) are hosted on Hugging Face… argh, it’s getting too technical, just read the link to see how it works!
https://huggingface.co/blog/claude-and-mcp
