"author:ki11a-Conton" matching MCP servers:

Enables the DeepSeek Harness to create and edit images through the chatgpt2api backend, auto-routing each job to text-to-image or image-to-image and running batches concurrently with ratio- and resolution-aware sizing. Results are saved to a local output directory under collision-free names, and responses come back over SSE with paste-ready Markdown image references so long jobs survive Cloudflare tunnel timeouts.