# WaveSpeed

> WaveSpeed is a paid MCP server for design and media. Run any WaveSpeed image, video, audio or 3D model from an MCP client, quoting price first. It can run locally on your machine or as a hosted remote server.

- **Kind:** MCP server
- **Pricing:** Paid
- **Platforms:** API, CLI
- **Website:** https://wavespeed.ai
- **Repository:** https://github.com/wavespeedai/mcp-server
- **MCP Registry name:** ai.wavespeed/mcp
- **Categories:** design and media, AI models and LLM tooling
- **Listing updated:** 2026-11-14

## About

This server follows the live WaveSpeed catalog rather than a fixed list: `list_models` searches it, `get_model_schema` shows the exact inputs a model accepts, and `run_model` submits a job, rejecting keys the model does not take. A `@path` marker inside inputs uploads a local file and swaps in its hosted URL. The catalog is cached for an hour, so newly released models work the day they ship.

Five shortcut tools (`generate_image`, `edit_image`, `generate_video`, `edit_video`, `extend_video`) pick an endpoint for what you supply and fit the request to the model's live schema, defaulting to Seedream 5.5 for images and Seedance 2.5 for video. They report any changes in notes and can return a price estimate without submitting anything.

Long generations can take minutes to hours, so `run_model` returns a prediction id and the assistant polls `get_prediction`; the task keeps running server-side. Account tools show credit balance, run history, itemized billing and usage, and a delete tool clears finished runs after confirmation. It runs over stdio with an API key or as a remote endpoint at api.wavespeed.ai/mcp.

## Features

- **Live model catalog:** Search and call any current model without a hardcoded list.
- **Schema before running:** Read a model's real input schema, then submit matching input.
- **Price quotes:** Estimate cost for the same input before spending.
- **Image and video shortcuts:** Generate, edit and extend media with defaults picked for you.
- **Long-running jobs:** Submit and poll by prediction id without losing work on disconnect.
- **Usage and billing views:** Check balance, history, itemized charges and spend.

## Good for

- Generate an image from a prompt inside Claude Code
- Edit or restyle existing pictures with a chosen model
- Extend or edit a video clip
- Quote the cost of a render before running it
- Review last week's spend per model

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http mcp https://api.wavespeed.ai/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://api.wavespeed.ai/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "mcp": {
      "url": "https://api.wavespeed.ai/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "mcp": {
      "type": "http",
      "url": "https://api.wavespeed.ai/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "mcp": {
      "serverUrl": "https://api.wavespeed.ai/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "mcp": {
      "type": "http",
      "url": "https://api.wavespeed.ai/mcp"
    }
  }
}
```

## Tools

- `list_models`: Search the live catalog by text or modality
- `get_model_schema`: A model's real input schema
- `run_model`: Submit a model run and return output URLs
- `generate_image`: Image from a prompt
- `edit_image`: Edit, restyle or combine images
- `generate_video`: Video from a prompt, frames or references
- `edit_video`: Edit an existing video
- `extend_video`: Continue an existing video
- `get_price`: Cost estimate with no charge
- `get_balance`: Account credit balance
- `list_predictions`: Run history
- `list_billings`: Itemized charges and refunds
- `get_usage`: Spend and request counts
- `delete_predictions`: Delete finished runs
- `upload_file`: Local file to hosted URL
- `get_prediction`: Status and outputs of a run

## FAQ

### Do I need an API key?

An API key is needed, though it is optional when the wavespeed CLI is already logged in.

### How is it priced?

WaveSpeed is pay-per-use with no monthly fee, and eligible new accounts get $1 in free credits.

### Can I preview cost?

Yes. `get_price` and `estimate_only` return an estimate without charging.

### Is there a remote option?

Yes, a Streamable HTTP endpoint at https://api.wavespeed.ai/mcp.

### Is WaveSpeed free?

No. WaveSpeed is a paid product.

### How do I add WaveSpeed to Claude Code?

Run this command in your terminal: claude mcp add --transport http mcp https://api.wavespeed.ai/mcp

### What are the best alternatives to WaveSpeed?

The closest alternatives to WaveSpeed are HyperFrames by HeyGen, ffmpeg-skill, Claude Code Ultimate Guide, AvatarSay — Fictional Character & Talking Video MCP and ChartAnt.

## Alternatives

- [HyperFrames by HeyGen](https://ai-powered.com/mcp/hyperframes-by-heygen): HyperFrames by HeyGen is a freemium MCP server for design and media.
- [ffmpeg-skill](https://ai-powered.com/mcp/ffmpeg-skill): ffmpeg-skill is an open-source MCP server for design and media.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [AvatarSay — Fictional Character & Talking Video MCP](https://ai-powered.com/mcp/avatarsay-fictional-character-and-talking-video-mcp): AvatarSay — Fictional Character & Talking Video MCP is a freemium MCP server for design and media.
- [ChartAnt](https://ai-powered.com/mcp/chartant): ChartAnt is a freemium MCP server for design and media.
- [VfxSmith](https://ai-powered.com/mcp/vfxsmith): VfxSmith is a freemium MCP server for design and media.
- [CalderFlow](https://ai-powered.com/mcp/calderflow): CalderFlow is a free MCP server for design and media.
- [Hourlore](https://ai-powered.com/mcp/hourlore): Hourlore is a freemium MCP server for design and media.

---
Source: https://ai-powered.com/mcp/wavespeed
