fastcar-vision
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| VISION_MODEL | No | Override the model name for the visual model. | |
| VISION_FRAMES | No | Default number of frames for video analysis (max 16). | 6 |
| VISION_API_KEY | No | Override the API key for the visual model. | |
| VISION_PROFILE | No | The visual model profile to use for this run. | |
| VISION_BASE_URL | No | Override the API Base URL for the visual model. | |
| MCP_VISION_CONFIG | No | Path to the visual configuration file. | ~/.mcp-vision-tools.json |
| VISION_TIMEOUT_MS | No | Total timeout for model and media in milliseconds. | 120000 |
| VISION_FFMPEG_PATH | No | Custom FFmpeg executable path. | ffmpeg-static |
| VISION_IMAGE_MODEL | No | Override the model name for image generation. | |
| MCP_VISION_STATE_DIR | No | Directory for daemon state and capability cache. | |
| VISION_IMAGE_API_KEY | No | Override the API key for image generation. | |
| VISION_IMAGE_PROFILE | No | Current image generation profile. | |
| VISION_IMAGE_BASE_URL | No | Override the API Base URL for image generation. | |
| MCP_VISION_IMAGE_CONFIG | No | Path to the image generation configuration file. | ~/.mcp-vision-tools.images.json |
| VISION_IMAGE_OPERATIONS | No | Comma-separated list of image operations to enable (generate,edit). | |
| VISION_IMAGE_TIMEOUT_MS | No | Total timeout for image generation operation in milliseconds. | 300000 |
| VISION_IMAGE_CONCURRENCY | No | Maximum concurrency for image generation requests (1-256). | |
| VISION_VIDEO_CONCURRENCY | No | Maximum concurrency for video analysis requests (1-256). | |
| VISION_ANALYSIS_QUEUE_LIMIT | No | Maximum number of analysis tasks in the queue (0-4096). | |
| VISION_IMAGE_ADAPTER_MODULE | No | Absolute path to a custom image provider adapter module (.mjs or .js). | |
| VISION_PROVIDER_CONCURRENCY | No | Maximum concurrency for provider requests (1-256). | |
| VISION_GENERATION_CONCURRENCY | No | Maximum concurrency for image generation tasks (1-32). | 2 |
| VISION_GENERATION_QUEUE_LIMIT | No | Maximum number of generation tasks in the queue (0-256). | 8 |
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 | 分析本地图片、file URL 或 HTTP(S) 图片。调用方决定 intent、图片处理模式和推理程度；支持流式进度、OCR、摘要和主体定位。 |
| analyze_videoA | 抽取本地或 HTTP(S) 视频帧并进行结构化理解，可指定帧数和模型 profile。 |
| list_vision_modelsA | 列出已配置的视觉模型 profiles 及当前默认项，不返回 API key。 |
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 3 tools
Each tool targets a clearly distinct purpose: image analysis, video analysis, and model listing. There is no overlap or ambiguity between them.
All tool names follow the consistent verb_noun pattern: analyze_image, analyze_video, list_vision_models. This provides a predictable and clear naming convention.
With 3 tools, the server is well-scoped and focused on its core capabilities. Each tool earns its place and the count is within the ideal range.
The server covers the primary functions of image/video analysis and model discovery. Minor gaps exist such as model management or more granular analysis options, but core workflows are well supported.