vision-mcp-server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| ZAI_API_KEY | No | Zhipu API key. | |
| VISION_DEBUG | No | Set to 'true' to output redacted debug logs to stderr. | false |
| VISION_MODELS | No | Comma-separated ordered list of vision model IDs for Zhipu or OpenAI-compatible providers. | |
| VISION_ROUTES | No | JSON string configuring routes for multiple providers. Overrides all simplified configuration. | |
| OPENAI_API_KEY | No | API key for the OpenAI-compatible API. | |
| OPENAI_BASE_URL | No | Base URL for the OpenAI-compatible API, e.g. https://provider.example/v1 | |
| VISION_PROVIDER | No | Provider selector for simplified configuration. Set to 'zhipu' for Zhipu or 'openai-compatible' for a generic OpenAI-compatible API. | |
| MODELSCOPE_MODEL | No | Legacy single model ID for ModelScope. Superseded by MODELSCOPE_MODELS if both are set. | |
| MODELSCOPE_TOKEN | No | ModelScope API token for the ModelScope provider. | |
| MODELSCOPE_MODELS | No | Comma-separated ordered list of ModelScope model IDs. All models share the same MODELSCOPE_TOKEN. | |
| VISION_API_KEY_ENV | No | Alternative environment variable name that contains the API key for simplified configuration. | |
| VISION_ALLOWED_DIRS | No | Comma-separated list of allowed local directories for image reading. | |
| VISION_MAX_IMAGE_EDGE | No | Maximum image edge length in pixels for simplified config. | 2048 |
| VISION_MAX_IMAGE_BYTES | No | Maximum image size in bytes for input or processed images. | 20971520 |
| VISION_MAX_IMAGE_PIXELS | No | Maximum pixel count for decoded images. | 40000000 |
| VISION_REQUEST_TIMEOUT_MS | No | Timeout in milliseconds for model API requests. | 60000 |
| VISION_FALLBACK_COOLDOWN_MS | No | Cooldown period in milliseconds for failed routes. | 60000 |
| VISION_IMAGE_DOWNLOAD_TIMEOUT_MS | No | Timeout in milliseconds for downloading online images. | 15000 |
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 | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| analyze_imageA | 使用可配置的视觉模型分析图片；遇到限流、超时或服务故障时可自动切换备用模型 |
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
Only one tool exists, so there is no possibility of confusion between tools. The name 'analyze_image' clearly describes its purpose.
With a single tool, naming consistency is trivially satisfied. The name 'analyze_image' follows a clear verb_noun pattern.
The server is named 'vision-mcp-server' implying a broad scope for vision tasks, but only one tool is provided. This is too few for the apparent scope, which would typically include at least model management and multiple analysis capabilities.
The tool covers basic image analysis but lacks supporting operations such as listing available models, batch processing, or other common vision tasks. These gaps can be worked around to some extent but are notable.