v1.1.0

Universal vision-language MCP server - image understanding & OCR for any MCP client.

Highlights:

• Model-agnostic: any OpenAI-compatible vision backend (Ollama native / vLLM / LM Studio / remote)
• Two tools: describe_image + ocr_image (multi-image, path-aware)
• Zero-dependency single-file executable - no Node/Python needed on target machines
• VLM_API_MODE dual protocol: ollama | openai
• Configurable via env: VLM_API_MODE / VLM_BASE_URL / VLM_MODEL / VLM_KEEP_ALIVE / VLM_TIMEOUT_MS

Assets: �ision-mcp.exe (Windows, single-file, ~89 MB)