deepseek-vision-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| VISION_API_KEY | Yes | Vision API key. | |
| VISION_TIMEOUT | No | Per-request timeout in seconds. | 60 |
| VISION_CACHE_DIR | No | Cache directory override (empty = platform default). | |
| VISION_OCR_MODEL | No | Model for mode=ocr. | |
| VISION_RETRY_MAX | No | Maximum retries on transient errors. | 2 |
| VISION_FULL_MODEL | Yes | Model for mode=full. | |
| VISION_CACHE_SCOPE | No | Cache scope: global, project, or none. | global |
| VISION_API_BASE_URL | Yes | OpenAI-compatible Chat API base URL. | |
| VISION_OCR_ENDPOINT | No | PaddleOCR dedicated OCR endpoint (full URL). Set when using PaddleOCR. | |
| VISION_CACHE_ENABLED | No | Cache switch. | true |
| VISION_CACHE_TTL_DAYS | No | Cache TTL in days. | 30 |
| VISION_FALLBACK_ENABLED | No | Fallback switch. | false |
| VISION_CACHE_MAX_SIZE_MB | No | Maximum cache size in MB. | 128 |
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
} |
| logging | {} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| extensions | {
"io.modelcontextprotocol/ui": {}
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| analyze_imageA | 当用户提供图片、截图、图表、照片，或要求查看图像内容时，调用本工具获取图片的视觉内容描述。 注意：本工具面向「自身无法直接查看图片」的主模型。若你自身具备原生视觉、能直接查看图片内容，应优先使用自己的原生视觉能力，无需调用本工具；仅在自身无法直接看图时才使用本工具。 默认使用 mode=full 进行完整视觉理解。仅当用户明确要求精确提取图片文字/表格内容时，使用 mode=ocr。 |
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
There is only one tool, so there is no possibility of confusing it with another tool. Its purpose is clearly scoped to image analysis.
The single tool name follows a clear verb_noun pattern and there are no conflicting naming conventions to cause inconsistency.
One tool is at the low end, but it is justified for a narrow vision-analysis server. The built-in full/OCR modes keep it reasonably well-scoped rather than feeling empty.
For the stated purpose of providing image understanding to a non-vision model, the tool covers both general visual description and explicit OCR text/table extraction. There are no obvious missing core operations for this narrow domain.