Codex vision bridge for DeepSeek V4 Flash: give text-only DeepSeek image capability in Codex. Local proxy turns pasted images and view_image into text via free GLM-4V-Flash or any OpenAI-compatible vision API. No GPU, no Ollama.
License-normalized reference retrieval for creative tools and agents — search images, video, audio, and text across pluggable sources via TypeScript SDK or MCP.
Run a close-to-real Claude Code session programmatically via the Agent SDK and expose it to a host app as something it can embed, watch, and control. Runner + wire protocol + gateway + embeddable panel.