A multimodal visual understanding MCP server based on the visual primitives paradigm. Through task scheduling, it separates scene understanding and coordinate localization into independent tools, enabling precise spatial reasoning: first see clearly, then locate.
Latest release v1.5.0 · 28 Jul 2026
2 files for Codex, OpenCode and Claude Code: visual-primitives-mcp AGENTS.md, visual-primitives-mcp CLAUDE.md — 4,740 tokens loaded in every session.
AGENTS.md A 4,512 tok CLAUDE.md A 228 tok These files are Wjl1224734792/visual-primitives-mcp's own configuration — they tell Codex, OpenCode and Claude Code how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.