Zerolab MCP Toolbox Server

You deploy and run an MCP server that lets a large language model interact with external services and APIs through a stable command interface. This enables you to perform tasks like executing system commands, accessing Figma data, reading and writing files, querying the web, generating images, and managing memory, all within an MCP-enabled client. This guide shows practical steps to install, configure, and use the Toolbox MCP server to extend your LLM capabilities.

How to use

You run the MCP server locally and connect your MCP-enabled client to it. Start the server using the standard stdio transport to have direct control from your LLM workflow. You can then issue commands, fetch data from Figma, read files, and perform web or media tasks through the model’s prompts.

How to install

Prerequisites: you need a modern shell, Node.js for some install options, and access to a package manager or runtimes described below.

# Install uv (recommended) on macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Or install UV via PowerShell on Windows
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Configuration and startup

To run the Toolbox MCP server locally with the complete feature set, start the server using stdio transport. The following command uses the all-feature set, including audio and memory tools.

uvx "mcp-toolbox[all]@latest" stdio

Using with an MCP client

Create a configuration that exposes the Toolbox MCP server to your MCP client. You can point the client at a local stdio server or a remote HTTP endpoint if provided.

The Toolbox can be configured with API keys for integrations such as Figma, Tavily, DuckDuckGo, and Flux image generation. These keys are supplied as environment variables in the server configuration.

Example environment-driven startup configuration blocks the required keys to enable full capabilities.

Available tools

execute_command

Execute a command line instruction from within the MCP-enabled client.

read_file_content

Read and return the contents of a file.

write_file_content

Write content to a specified file.

replace_in_file

Replace content in a file using regular expressions.

list_directory

List the contents of a directory with detailed information.

figma_get_file

Retrieve a Figma file by key.

figma_get_file_nodes

Fetch specific nodes from a Figma file.

figma_get_image

Retrieve images for nodes in a Figma file.

figma_get_image_fills

Get URLs for images used in a Figma file.

figma_get_comments

Get comments on a Figma file.

figma_post_comment

Post a comment on a Figma file.

figma_delete_comment

Delete a comment from a Figma file.

figma_get_team_projects

Get projects for a team in Figma.

figma_get_project_files

Get files for a project in Figma.

figma_get_team_components

Get components for a team.

figma_get_file_components

Get components from a file.

figma_get_component

Get a component by key.

figma_get_team_component_sets

Get component sets for a team.

figma_get_team_styles

Get styles for a team.

figma_get_file_styles

Get styles from a file.

figma_get_style

Get a style by key.

xiaoyuzhoufm_download

Download a XiaoyuZhouFM podcast episode with optional conversion.

get_audio_length

Return the duration of an audio file in seconds.

get_audio_text

Transcribe audio text for a given time range.

think

Append a memory-like thought to the log for later recall.

get_session_id

Retrieve the current session ID.

remember

Store a memory in the memory database.

recall

Query memories using semantic search from the memory database.

forget

Clear all memories from the memory database.

convert_file_to_markdown

Convert a file to Markdown using MarkItDown.

convert_url_to_markdown

Convert a URL to Markdown using MarkItDown.

get_html

Fetch HTML content from a URL.

save_html

Save HTML content from a URL to a file.

search_with_tavily

Search the web using Tavily (requires API key).

search_with_duckduckgo

Search the web using DuckDuckGo (requires API key).

flux_generate_image

Generate an image via the Flux API and save it to a file.