RunAPI Vidu MCP Server
<h1 align="center">RunAPI Vidu MCP Server</h1>
<p align="center">
<strong>Vidu API access for AI agents: run video generation operations, poll asynchronous results, and check pricing through one focused MCP server.</strong>
</p>
<p align="center">
<sub>Works with Claude Code, Codex, Cursor, Windsurf, VS Code, Roo Code, and any MCP-compatible host.</sub>
</p>
<p align="center">
<a href="https://www.npmjs.com/package/@runapi.ai/vidu-mcp"><img src="https://img.shields.io/npm/v/%40runapi.ai/vidu-mcp?style=flat-square&color=blue" alt="npm version"></a>
<a href="https://github.com/runapi-ai/vidu-mcp"><img src="https://img.shields.io/badge/GitHub-runapi--ai%2Fvidu--mcp-24292f?style=flat-square" alt="GitHub repository"></a>
<a href="LICENSE"><img src="https://img.shields.io/badge/License-Apache_2.0-blue?style=flat-square" alt="Apache-2.0 license"></a>
<img src="https://img.shields.io/badge/Type-MCP_Server-blue?style=flat-square" alt="MCP Server">
<img src="https://img.shields.io/badge/Models-2-16a34a?style=flat-square" alt="2 models">
</p>
<p align="center">
<a href="#install">Install</a> |
<a href="#tools">Tools</a> |
<a href="#models">Models</a> |
<a href="#agent-prompts">Agent Prompts</a> |
<a href="#configuration">Configuration</a> |
<a href="#links">Links</a>
</p>
---
## Why This Package?
`@runapi.ai/vidu-mcp` is a focused Model Context Protocol server for the **Vidu** model line on RunAPI.
It gives MCP-compatible assistants direct access to 2 endpoints and 2 model variants without loading the full RunAPI catalog.
Use this per-model server when an agent should stay scoped to Vidu. Use [`@runapi.ai/mcp`](https://github.com/runapi-ai/mcp) when one assistant should discover every RunAPI model line.
---
## Install
Add it to Claude Code:
```bash
claude mcp add vidu -s user -- npx -y @runapi.ai/vidu-mcp
```
Use project scope when the server should be shared with a repository:
```bash
claude mcp add vidu -s project -- npx -y @runapi.ai/vidu-mcp
```
Codex, Cursor, Windsurf, VS Code, Roo Code, and other MCP hosts can use the same stdio command:
```json
{
"mcpServers": {
"vidu": {
"command": "npx",
"args": ["-y", "@runapi.ai/vidu-mcp"]
}
}
}
```
`check_pricing` works before sign-in. For task creation and status polling, ask your assistant to call the `login` tool. It opens a browser login and saves credentials to `~/.config/runapi/config.json`, the same file used by `runapi login`.
Headless and CI hosts can still set `RUNAPI_API_KEY` before starting the MCP host.
Ready-made examples are in [`examples/`](examples/) for Claude, Cursor, Windsurf, VS Code, and Roo Code.
---
## Tools
| Tool | Auth | Purpose |
|---|---|---|
| `image_to_video` | Yes | Create a Vidu image to video task and optionally wait for a terminal status. Returns the task id, status, and output URLs. |
| `reference_to_video` | Yes | Create a Vidu reference to video task and optionally wait for a terminal status. Returns the task id, status, and output URLs. |
| `get_task` | Yes | Fetch the current status and latest payload for an existing task. |
| `check_pricing` | No | Look up current pricing for a Vidu model and endpoint. |
---
## Models
Vidu covers 2 model variants across 2 endpoints. Each tool accepts the models listed for it:
| Tool | Models |
|---|---|
| `image_to_video` | `vidu-q4-image-to-video` |
| `reference_to_video` | `vidu-q4-reference-to-video` |
Model availability can change between releases. Use `check_pricing` or the [Vidu model page](https://runapi.ai/models/vidu) for the current catalog view.
---
## Agent Prompts
Ask your assistant in natural language; it can inspect pricing, create the task, and return the task id plus output URLs.
### Create a task
```text
Run a Vidu image to video task with RunAPI.
```
The assistant can call `check_pricing`, then `image_to_video`, and return the task id, status, and output URLs.
### Submit without waiting
```text
Create the task but don't wait for it to finish.
```
The assistant calls the create tool with `wait: false` and returns the task id. Check on it later with `get_task`.
### Check pricing before creating
```text
Check current Vidu pricing, then create the task if it matches my request.
```
The assistant calls `check_pricing` and can link to the [Vidu model page](https://runapi.ai/models/vidu) for the canonical catalog entry.
---
## Configuration
The server resolves auth in this order:
1. `RUNAPI_API_KEY` environment variable, useful for headless and CI hosts
2. `~/.config/runapi/config.json`, created by the MCP `login` tool or `runapi login`
3. No key, which still allows `check_pricing`
The config file is normally managed by login. A pre-provisioned headless config can use:
```json
{
"apiKey": "your_runapi_key"
}
```
Do not commit real API keys.
---
## Links
| Resource | URL |
|---|---|
| Vidu model page | [https://runapi.ai/models/vidu](https://runapi.ai/models/vidu) |
| npm package | [@runapi.ai/vidu-mcp](https://www.npmjs.com/package/@runapi.ai/vidu-mcp) |
| GitHub repository | [runapi-ai/vidu-mcp](https://github.com/runapi-ai/vidu-mcp) |
| RunAPI MCP overview | [runapi.ai/mcp](https://runapi.ai/mcp) |
| RunAPI docs | [runapi.ai/docs](https://runapi.ai/docs) |
---
## License
Licensed under the [Apache License, Version 2.0](LICENSE).
TDQS
Scored across 5 tools
Each tool maps to a distinct purpose: login, task creation (two variants), task polling, and pricing. The image_to_video vs reference_to_video pair shares the same 'create a Vidu task' framing and could momentarily confuse, but the descriptions clearly distinguish the two input modes.
Most names follow a verb_noun convention (get_task, check_pricing) or a clear X_to_video pattern shared by the two generation tools. login is a bare verb that breaks the pattern slightly, but the set remains readable and predictable overall.
Five tools is well-scoped for a video-generation API: auth, two generation modes, polling, and pricing. Nothing feels redundant, though the surface is on the leaner side.
Core workflow (login, create task, poll task, check pricing) is covered, but there is no tool to list existing tasks, cancel a task, or access other plausible Vidu modes (e.g. text-to-video). get_task partially compensates but leaves dead ends for task management.