General and MCP Settings
Adjust theme, language, gateway port, API Key, AI Agent, and log retention settings
General and MCP Settings
The Settings page manages the global behavior of MCP2Skill. Although it does not create MCP services directly, much of the runtime behavior depends on the configuration here.
From a product perspective, this page manages the "unified runtime layer" of MCP2Skill:
- Run MCP in one place, instead of each client running its own copy
- Manage Skills installation targets in one place
- Centrally control the port, API Key, search paths, and log retention
This is also the key foundation that lets MCP2Skill save local machine resources and eliminate duplicated configuration costs.
Screenshot placeholder: overview of the Settings page, showing the left sub-navigation and the General, MCP, AI Provider, AI Agents, Data Management, and About sub-pages.
Video placeholder: a demo walking through the Settings page to adjust the theme, gateway port, API Key, AI Provider, AI Agent directories, and log retention count in sequence.
General page
The General page mainly covers application-level experience settings.
Common items in the current version include:
- Theme
- Language
- Translucent sidebar effect
- Launch at startup
- Minimize to tray
If you use MCP2Skill across different system environments, the theme and language are usually the first things to adjust.
MCP settings page
The MCP settings page is one of the most important global pages, and it mainly contains the following categories.
Gateway port
This determines which port external clients use when connecting to MCP2Skill.
After changing the port, it is recommended to:
- Re-copy the endpoint or JSON configuration
- Update the connection details in your external clients
- Run an actual call to verify
If you change the port but do not update your clients accordingly, the clients will usually fail to connect.
Remote access
If you want external environments to reach MCP2Skill over the network, pay attention to the "Remote Access" toggle.
Before enabling it, confirm that:
- Your use case actually requires remote access
- The current network environment is under your control
- You have considered whether to enable the API Key at the same time
If you only use MCP2Skill locally, you usually do not need to turn it on.
API Key
The API Key controls who can access the MCP2Skill gateway.
The current version supports:
- Enabling or disabling API Key authentication
- Viewing the current API Key
- Copying the API Key
- Regenerating the API Key
Once you regenerate the API Key, the configuration in your existing clients usually needs to be updated as well.
MCP search paths
MCP search paths are useful for adding extra directories that MCP2Skill should scan or recognize.
The current version supports:
- Adding a path
- Browsing directories
- Editing a path
- Reordering paths
- Deleting a path
Only add directory paths that actually exist and are stable.
AI Provider page
The AI Provider page configures the AI service that MCP2Skill calls when generating Skill content.
Common settings include:
- Provider type
- Base URL or compatible API endpoint
- API Key
- Default model
If Skill generation fails, returns empty responses, or the model is unavailable, in addition to checking the MCP service itself, come back here to verify that the AI Provider configuration is correct.
AI Agents page
If you want to install Skills, you first need to configure the target directories on the AI Agents page.
Here you can:
- Add an AI Agent
- Name an AI Agent
- Specify the Skill directory
- Edit existing configurations
- Delete targets that are no longer used
This step directly affects how Skills are installed and browsed.
Why unified management saves resources
If you configure the same set of MCP separately in multiple AI clients, you typically end up with:
- Each client maintaining its own copy of the configuration
- Each client starting its own processes
- Each client producing its own logs and debugging overhead
With MCP2Skill, you centralize these actions in one place, which usually reduces the overhead of duplicated startups, duplicated configuration, and duplicated maintenance.
Data Management page
The Data Management page is more maintenance-oriented. Common items in the current version include:
- Call log retention count
- Clear all logs
If you are debugging frequently, keep the log retention count a bit higher; if you care more about local data size, shrink it to fit your habits.
About page
The About page is mainly for checking:
- Application version
- Tauri version
- External reference links
This page is useful when you need to report an issue or verify version numbers.
Recommended settings order
If this is your first visit to the Settings page, we recommend going through it in this order:
- General
- MCP
- AI Provider
- AI Agents
- Data Management
- About
Common recommendations
- After changing the port, remember to update your clients accordingly
- When enabling remote access, consider enabling the API Key at the same time
- After regenerating the API Key, reconfigure your external clients
- Before generating Skills, verify that the AI Provider is available
- Before installing Skills, configure the AI Agent directories first
- To reduce duplicated local resource usage, let clients reuse MCP2Skill instead of each maintaining its own set of MCP
Related chapters
- To learn how clients connect: see Connect External AI Clients
- To install a Skill: see Generate and Install Skills
- To troubleshoot log issues: see Logs and Diagnostics