lawrencewzen/pi-vlm-proxy

`pi-deepseek-vision` is a vision preprocessing extension for Pi 0.84.1. It uses a vision model already available in Pi to analyze images in the current visible context, replaces them with numbered markers and plain-text analysis, and then calls the target DeepSeek model. The DeepSeek model itself always remains text-only.

Give text-only LLMs vision. A tiny OpenAI-compatible proxy that lets reasoning models (DeepSeek, Qwen, GLM…) see images by querying a separate vision model through tools: look, OCR, scan, crop, compare. No training, no weights.

VisionPower: portable image understanding MCP server for Codex, Claude Desktop, Cursor, and other agents.

CLIProxyAPI v7 plugin for DeepSeek Responses vision preprocessing

Eyes for text-only DeepSeek Harness agents: built-in free vision chain (no key) + pixel-level vision tools (Q&A, grounding, crop, pixel diff, colors, OCR, SVG trace, cutout, screenshots). One-command install, no Python, image turns work like ordinary tool-calling turns.

A token-efficient, provider-agnostic vision layer for reasoning models via MCP.

DeepSeek-VL: Towards Real-World Vision-Language Understanding

give non-vision models the ability to see images by routing images to a vision capable model