Skip to main content
Glama
ViralFaceVideo Template Finder

ViralFaceVideo Template Finder

README.md
# ViralFaceVideo Template Finder

Find a photo-to-video effect for a short dance, celebration, or rap clip. Compare how many photos each template needs, its available lengths, and its credit costs. Preview your chosen template and generate the clip on the [ViralFaceVideo website](https://viralfacevideo.com).

## MCP

Install `viralfacevideo.mcpb` in an MCPB-compatible desktop client. For a client with stdio support, clone this repository, run `npm ci` with Node.js 20 or later, and configure this server:

```json
{
  "mcpServers": {
    "viralfacevideo": {
      "command": "node",
      "args": ["/absolute/path/to/viralfacevideo-agent/src/index.mjs"]
    }
  }
}
```

Ask, for example, “Find a rap template for two photos and show me the shortest option.” The `find_video_templates` tool supports `photos` (1 or 2) and `query` (English keywords). It fetches the [public catalog feed](https://viralfacevideo.com/llms.txt) and returns official preview links. No API key is required. Photos stay with you until you choose to upload them on the website.

## Agent Skill

From a checkout, list the skill with `npx skills add . --list`. Once the public repository is available, install it with:

```sh
npx skills add sakiila/viralfacevideo-agent --skill viralfacevideo-templates
```

The [skill](skills/viralfacevideo-templates/SKILL.md) also works independently by reading the public catalog feed when MCP is unavailable.

## 验证和发布

1. 运行 `npm ci`、`npm test`、`npm run smoke`。冒烟测试通过 stdio 调用 MCP,并读取官网实时模板目录。
2. 用 `npm run bundle` 生成 `.tmp/viralfacevideo.mcpb`,需要系统安装 `zip`。如果重新打包,用 `shasum -a 256 .tmp/viralfacevideo.mcpb` 更新 `server.json` 里的 `fileSha256`,确保登记的哈希对应实际上传的文件。
3. 将这个目录作为独立的公开仓库 `sakiila/viralfacevideo-agent` 发布,并把 MCPB 附到与 `server.json` 一致的 GitHub Release。公开仓库可供 Skills CLI 安装。
4. Release 可匿名下载后,运行 `mcp-publisher login github` 和 `mcp-publisher publish`,登记到官方 MCP Registry。验证登记返回的 `websiteUrl` 指向官网。[MCP Registry 发布文档](https://modelcontextprotocol.io/registry/quickstart)

TDQS

B3.4/5.0

Scored across 1 tool

Disambiguation5/5

There is only a single tool, so there is no possibility of confusing it with another operation. Its purpose—searching templates by photo count and keywords—is clearly stated.

Naming Consistency4/5

The lone name follows a predictable verb_noun snake_case pattern (find_video_templates) that reads naturally. With just one tool there is no cross-tool naming convention to verify, keeping it slightly below a perfect score.

Tool Count3/5

One tool is defensible for a narrowly scoped template-search service, but the surface feels thin and leaves no room for related lookups. It sits at the borderline of the 1-2 tool range.

Completeness3/5

Search is covered and returns useful metadata (preview pages, durations, credit costs), but there are notable gaps: no detail/retrieval by template ID, no category or duration browsing, and no way to enumerate available templates. Agents can work around this only partially.

Maintenance

ActivityMaintained
ResponsivenessNo recent community issues