# vision-gemma

> vision-gemma — q1023884985-visiongemma. Use this tool when you need to enable AI assistants with offline image recognition, OCR, and scene description capabilities while maintaining user privacy. It solves problems related to visual data processing, text extraction, and scene understanding in resource-constrained or secure environments. The tool accepts image inputs and outputs recognized text, objects, and scene descriptions via the Model Context Protocol.

Canonical page: https://skillsregistry.net/skills/q1023884985-visiongemma  
JSON: https://api.skillsregistry.net/v1/skills/q1023884985-visiongemma

## Description

Provides offline, privacy-preserving image recognition, OCR, and scene description for AI assistants via the Model Context Protocol, with Vulkan-accelerated local processing.

## Trust

- **Trust score (0–1):** 0.60
- **Verification tier:** scanned
- **Last scanned:** 2026-09-03

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** media
- **Updated:** 2026-09-03

## Source

- **Source listing:** [Glama](https://glama.ai/mcp/servers/o7u07lqr7v)
- **Repository:** <https://github.com/q1023884985/visiongemma>

## 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": "q1023884985-visiongemma"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/q1023884985-visiongemma` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/q1023884985-visiongemma/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