Skip to main content
Glama
Sign In
TensorsLab MCP Server

TensorsLab MCP Server

by bob5-tensorslab
OverviewSchemaSourceRelated ServersDiscussionsAdmin
JavaScript
Remote

TensorsLab MCP Server

A Model Context Protocol (MCP) server that provides tools for generating AI images and videos using the TensorsLab API.

Features

Image Generation Tools

  • generate_image_v45 - SeeDream V4.5 (highest quality, recommended)

  • generate_image_v4 - SeeDream V4 (faster generation)

  • check_image_task_status - Check image generation status

  • delete_image_tasks - Delete image tasks

Video Generation Tools

  • generate_video_v2 - SeeDance V2 (latest, up to 15s, 1440p)

  • generate_video_v15pro - SeeDance V1.5 Pro (high quality)

  • generate_video_fast - SeeDance V1 Fast (quick previews)

  • check_video_task_status - Check video generation status

  • delete_video_tasks - Delete video tasks

Related MCP server: MCP Veo 3 Video Generation Server

Installation

1. Install Dependencies

cd mcp/tensorslab-mcp-server
npm install

2. Build the Project

npm run build

Configuration

Getting Your API Key

  1. Visit https://test.tensorai.tensorslab.com/

  2. Sign up or log in

  3. Navigate to API settings

  4. Copy your API key

Claude Desktop Configuration

Add the following to your Claude Desktop config file:

Windows: %APPDATA%\Claude\claude_desktop_config.json macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "tensorslab": {
      "command": "node",
      "args": ["tensorslab-mcp-server/dist/index.js"],
      "env": {
        "TENSORSLAB_API_KEY": "your-api-key-here"
      }
    }
  }
}

Important: Replace your-api-key-here with your actual TensorsLab API key and update the path if needed.

Usage

Image Generation

// Generate an image with SeeDream V4.5
await generate_image_v45({
  prompt: "A beautiful sunset over mountains",
  batchSize: 2,
  resolution: "16:9"
})

Video Generation

// Generate a video with SeeDance V2
await generate_video_v2({
  prompt: "A drone flying over a forest",
  ratio: "16:9",
  duration: 10,
  resolution: "1080p"
})

Tool Parameters Reference

Image Generation (SeeDream V4.5)

Parameter

Type

Required

Description

prompt

string

Yes

Text description of the image

batchSize

number

No

1-15 images (default: 1)

resolution

string

No

9:16, 16:9, 3:4, 4:3, 1:1, 2:3, 3:2, 2K, 4K, or WxH

imageUrl

string

No

Source image URL for image-to-image

seed

number

No

Random seed for reproducibility

promptExtend

boolean

No

Enable prompt enhancement (default: true)

Video Generation (SeeDance V2)

Parameter

Type

Required

Description

prompt

string

Yes

Text description of the video

ratio

string

No

9:16 (vertical) or 16:9 (horizontal)

duration

number

No

5-15 seconds (default: 5)

resolution

string

No

480p, 720p, 1080p, or 1440p

fps

string

No

Frame rate (default: 24)

imageUrl

string

No

Source image URL for image-to-video

generateAudio

boolean

No

Generate audio (default: false)

returnLastFrame

boolean

No

Return last frame as image

seed

number

No

Random seed for reproducibility

Development

Run in Development Mode

npm run dev

Run with MCP Inspector

npm run inspector

Build for Production

npm run build
npm start

Project Structure

tensorslab-mcp-server/
├── src/
│   ├── index.ts      # Main MCP server with tool registrations
│   ├── api.ts        # TensorsLab API client
│   └── types.ts      # TypeScript type definitions
├── dist/             # Compiled JavaScript (generated)
├── package.json
├── tsconfig.json
└── README.md

Troubleshooting

"TENSORSLAB_API_KEY environment variable is required"

  • Make sure you set the API key in your Claude Desktop config or environment

"Insufficient credits"

Task timeout

  • Large videos may take longer. Use check_video_task_status to poll manually

  • Image tasks timeout after 3 minutes, video tasks after 5 minutes

Module not found errors

  • Run npm install to ensure all dependencies are installed

  • Run npm run build to compile TypeScript

API Reference

For complete API documentation, see:

License

MIT

Available Tools

9 tools
check_image_task_statusC

Check the status of an image generation task

Inputs: taskId (required).

check_video_task_statusC

Check the status of a video generation task

Inputs: taskId (required).

delete_image_tasksC

Delete one or more image generation tasks

Inputs: taskIds (required).

delete_video_tasksC

Delete one or more video generation tasks

Inputs: taskIds (required).

generate_image_v4C

Generate an image using TensorsLab SeeDream V4 model (faster generation with good quality)

Inputs: prompt (required), batchSize, resolution, imageUrl.

generate_image_v45B

Generate an image using TensorsLab SeeDream V4.5 model (highest quality, recommended for general use)

Inputs: prompt (required), batchSize, resolution, imageUrl, +2 more.

generate_video_fastB

Generate a video using TensorsLab SeeDance V1 Fast model (faster generation, good for previews)

Inputs: prompt (required), ratio, duration, resolution, +1 more.

generate_video_v15proC

Generate a video using TensorsLab SeeDance V1.5 Pro model (high quality, 5-10 seconds, up to 1080p)

Inputs: prompt (required), ratio, duration, resolution, +1 more.

generate_video_v2B

Generate a video using TensorsLab SeeDance V2 model (highest quality, supports 5-15 seconds, up to 1440p)

Inputs: prompt (required), ratio, duration, resolution, +5 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    B3.3/5.0

    Scored across 9 tools

    Disambiguation3/5

    The tools are grouped into clear categories (image vs. video, generation vs. status check vs. deletion), but there is significant overlap within generation tools. For example, generate_image_v4 and generate_image_v45 differ only by model version, which could cause confusion about which to use. Similarly, the three video generation tools have overlapping purposes with quality/speed trade-offs that might not be immediately clear to an agent.

    Naming Consistency5/5

    All tool names follow a consistent snake_case pattern with a clear verb_noun structure. The naming is highly predictable: check_<resource>_task_status, delete_<resource>_tasks, and generate_<resource>_<model> for both image and video resources. This consistency makes it easy for an agent to understand and navigate the tool set.

    Tool Count4/5

    With 9 tools, the count is reasonable for a media generation server. It covers core operations (generate, check status, delete) for two resource types (image and video), which is well-scoped. However, it might be slightly heavy due to multiple model-specific generation tools, but this is justified by the domain's need for different quality/speed options.

    Completeness4/5

    The tool set provides good coverage for image and video generation workflows, including creation (generate), status monitoring (check), and cleanup (delete). Minor gaps exist, such as no tools for updating tasks or retrieving generated media directly, but agents can work around these using the provided status and deletion tools in typical workflows.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Resources

    Related MCP Connectors

    Related MCP Servers