# Visual Memory MCP

This server stores visual UI states and helps an agent find elements or compare screens. Use it to inspect interface changes and trace recorded tests.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 39
- Author: putervision
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/putervision/vision-memory-mcp
- Updated: 2026-10-08
- HTML: https://affpapa.org/en/nevibecode/mcp/visual-memory-mcp

## What it does

Visual Memory MCP keeps interface states locally and searches for similar screens using perceptual hashes and local CLIP vectors. An agent can retrieve states from text or images, map interface elements to CSS selectors and coordinates, and trace transitions between screens.

The server processes WebM and MP4 recordings, extracts keyframes, and links them to state transitions. You can also register a screenshot or mockup as a baseline for visual regression checks and save state snapshots. Node.js 18.18.0 or later is required. A startup option skips the CLIP model download, but semantic search relies on that model.

## Requirements
- Needs: Node.js (commands run via npx)

## Install

### Claude Code (native platform)

**Platform CLI**
```bash
claude mcp add vision-memory-mcp -- npx -y @putervision/vision-memory-mcp
```

### Codex

**Platform CLI**
```bash
codex mcp add visual-memory-mcp -- npx -y @putervision/vision-memory-mcp
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "visual-memory-mcp": {
            "command": "npx",
            "args": [
                "-y",
                "@putervision/vision-memory-mcp"
            ]
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "visual-memory-mcp": {
            "type": "stdio",
            "command": "npx",
            "args": [
                "-y",
                "@putervision/vision-memory-mcp"
            ]
        }
    }
}
```

### Gemini CLI

**Platform CLI**
```bash
gemini mcp add visual-memory-mcp npx -y @putervision/vision-memory-mcp
```

## When to use it
- find a saved screen from an image or description
- map an interface element to a CSS selector and coordinates
- trace screen changes in a WebM or MP4 recording
- compare the UI against a registered mockup
- save and restore a visual state snapshot

## Example prompts
- Find a saved state that resembles this screenshot with `recall_memory`.
- Locate the button and return its selector and coordinates using `predict_next_action`.
- Analyze `./playwright-test.webm` and trace the screen transitions.

## Similar
- [REA](https://affpapa.org/en/nevibecode/mcp/rea.md) — MCP server · ★ 386989
- [GitHub](https://affpapa.org/en/nevibecode/mcp/github.md) — MCP server · ★ 33918
- [Netdata](https://affpapa.org/en/nevibecode/mcp/netdata.md) — MCP server · ★ 81115
- [browser-use](https://affpapa.org/en/nevibecode/mcp/browser-use.md) — MCP server · ★ 119160
- [Codebase Memory](https://affpapa.org/en/nevibecode/mcp/codebase-memory.md) — MCP server · ★ 47392
- [mcp-server](https://affpapa.org/en/nevibecode/mcp/mcp-server.md) — MCP server · ★ 43872

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md