Sn Infographic

The main agent handles parameter extraction while a worker agent runs the pipeline: evaluates your prompt quality, optionally expands it, generates images across multiple rounds with VLM review, then ranks results. You get quality-based output in friendly mode (best image only) or verbose mode (full stats and all ranked images). It pulls from the sn-image-base toolkit and supports 1-8 generation rounds with early termination when quality thresholds hit. Useful when you need actual visual summaries instead of just charts. The multilingual triggers and aspect ratio inference show it's built for production use, though the SenseNova API dependency means you'll need their keys to run it.