seevido
# SeeVido Prompt MCP
A deterministic, read-only MCP server for English or Chinese video prompts, reference-aware sequences, diagnosis, and canonical [SeeVido](https://seevido.org) resources. It does not generate media, call model providers, access accounts, compare live prices, or spend credits.
## Tools
- `build_video_prompt`
- `plan_reference_sequence`
- `diagnose_video_prompt`
- `get_seevido_resources`
All tools declare read-only MCP annotations and require no API key.
## Install
```json
{
"mcpServers": {
"seevido": {
"command": "npx",
"args": ["-y", "github:gpt-img-2/seevido-prompt-mcp"]
}
}
}
```
Or run locally with `pnpm install`, `pnpm build`, and `node dist/index.js`. Set `SEEVIDO_APP_BASE_URL` only when canonical resource links should use another origin.
The companion OpenClaw Skill is at [`openclaw/seevido-video-prompt-architect`](openclaw/seevido-video-prompt-architect/SKILL.md). This is an independent prompt-design utility, not official documentation for any model provider.
## Development
```bash
pnpm validate
```
## License
MIT
TDQS
Scored across 4 tools
Each tool has a distinct purpose: building prompts, planning sequences, diagnosing prompts, and retrieving resources. There is no overlap in functionality, so an agent can clearly identify which tool to use for a given task.
All tool names follow the verb_noun pattern in snake_case (build_video_prompt, plan_reference_sequence, diagnose_video_prompt, get_seevido_resources). The verbs clearly describe the action and the nouns the target, making the naming predictable and uniform.
With exactly 4 tools, the server is well-scoped and each tool earns its place for the domain of video prompt engineering. This is within the ideal 3-15 range and not excessive for the functionality offered.
The tools cover the core lifecycle of video prompt creation: building, planning sequences, diagnosing, and obtaining resources. Minor gaps exist, such as no tool for editing or deleting existing prompts, but these are not essential for the server's apparent purpose and agents can work around them.