# Vision Web Bridge

> Use this tool when you need to analyze images or read web pages using OpenAI-compatible vision models, and integrate the results with MCP clients like Claude Desktop. It solves problems related to image analysis and web content summarization, accepting various image inputs and providing secure, localized processing. The Vision Web Bridge requires an OpenAI-compatible API endpoint and credentials, and is suitable for use cases where secure, automated image and web analysis is necessary.

Canonical page: https://skillsregistry.net/skills/dangpolly927-eng-vision-bridge  
JSON: https://api.skillsregistry.net/v1/skills/dangpolly927-eng-vision-bridge

## Description

Vision Web Bridge connects MCP clients like Claude Desktop to OpenAI-compatible vision models for image analysis and web reading. It supports images from uploads, clipboard, local paths, URLs, or base64 input. The server extracts and summarizes web page content locally, with security defaults that block private network access and disable clipboard by default. Requires an OpenAI-compatible API endpoint and credentials.

## Trust

- **Trust score (0–1):** 0.65
- **Verification tier:** scanned
- **Last scanned:** 2026-09-01

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** media
- **Updated:** 2026-09-01

## Source

- **Source listing:** [PulseMCP](https://www.pulsemcp.com/servers/dangpolly927-eng-vision-bridge)
- **Repository:** <https://github.com/dangpolly927-eng/mcp-vision-web-bridge>

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "dangpolly927-eng-vision-bridge"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/dangpolly927-eng-vision-bridge` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/dangpolly927-eng-vision-bridge/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt