vision-relay-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| VISION_MODEL | Yes | Model name for vision analysis, e.g., claude-sonnet-4-6 or gpt-4o. | |
| VISION_API_KEY | Yes | Your relay API key. | |
| VISION_BASE_URL | Yes | Your relay API base URL. | |
| VISION_PROVIDER | Yes | Selects the API format: 'anthropic' or 'openai'. | |
| VISION_MAX_TOKENS | No | Max tokens for the vision response. | 2000 |
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 |
|---|---|
| process_imagesA | General-purpose vision tool for one or more local images. Use it for OCR, extracting structured information, describing content, answering questions, locating elements, summarizing screenshots, or reasoning across multiple images in a single API call. |
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
With only one tool, there is no potential for confusion between tools. The tool's distinct purpose is clear.
A single tool name does not violate any naming patterns, and it is self-consistent.
One general-purpose tool can cover many vision tasks adequately, but a server dedicated to vision might benefit from having separate tools for distinct operations (e.g., OCR, Q&A) to improve clarity and agent selection.
The tool covers a wide range of vision tasks including OCR, extraction, description, Q&A, element location, summarization, and multi-image reasoning. Minor gaps exist (e.g., no video or generation), but the core domain is well-served.