dsh-vision-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| VISION_MODEL | No | Vision model name | gpt-4o-mini |
| VISION_API_KEY | No | API key; can be omitted for local Ollama and other key-free services | |
| VISION_API_BASE | No | OpenAI-compatible API base URL (without /chat/completions) | https://api.openai.com/v1 |
| VISION_PROVIDERS | No | JSON array of provider configurations for fallback | |
| VISION_REASONING | No | Global fallback reasoning effort: none, low, medium, high | |
| VISION_MAX_TOKENS | No | Maximum output tokens | 4096 |
| VISION_TIMEOUT_MS | No | Request timeout in milliseconds | 120000 |
| VISION_TEMPERATURE | No | Sampling temperature | 0.2 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| img2textA | 读取图片并把内容转换为文字描述。接受本地文件路径、http(s) URL、data URL 或 base64。由外部视觉模型（VISION_PROVIDERS 配置的一组 OpenAI 兼容模型，自动 fallback）完成识别，返回纯文本，供本身不支持图片输入的模型理解图片。使用时机：① 你能直接查看图片时（原生多模态输入），无需调用本工具；② 用户提供图片路径/URL/图片附件引用（形如 "[图片附件：名称] 图片文件：<绝对路径>"）且你无法直接看到图片内容时，立即调用本工具读取并转述内容，不要回复"无法查看图片"；③ 用户要求分析/读取某张图片时。prompt 参数：用户有具体问题时传用户的问题；没有问题时不要传（server 使用默认提示词）。转述时保持简洁：先一句话概括图片是什么，再列关键信息（主要文字、重点元素），不要输出大段分析。 |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 1 tool
The server exposes only one tool, so there is no possibility of confusing it with another tool. Its purpose—converting an image to text—is clearly stated and unambiguous.
With a single tool, there are no naming conflicts or mixed conventions to penalize. 'img2text' is a concise, descriptive name that clearly communicates image-to-text conversion.
One tool feels minimal for a vision-oriented MCP server, even though the tool is focused and non-trivial. It is borderline: functional for a narrow use case, but the overall surface is thin.
The single tool covers the core image-to-text workflow well, supporting local paths, URLs, data URLs, and base64 input. A broader vision server might include additional operations like structured extraction or image comparison, but for its stated purpose there are no major dead ends.