🚀 Quickstart
This is a quick guide to get started with QueryPie AI Hub.
By following this guide, you’ll be able to integrate various SaaS apps and use them in AI chats within just a few minutes.
Creating an AI Hub Account
First, let’s create an AI Hub account.
- Visit the QueryPie AI Hub homepage and click the Login button.
- We currently support Google, GitHub, and email registration.
- To sign up with Google or GitHub account, click the Continue with Google / Continue with GitHub button.
- To sign up with email, click the Sign up button below the email and password input fields, then enter your name, email, and password.
- Enter additional member information and complete the registration.
Currently, registration is only possible with invited emails.
If you need an invitation, please contact aihub@querypie.com.
Creating Integrations
Integration is a feature that connects with external services or databases.
For example, it allows you to use various services like Slack, JIRA, Notion through MCP.
Selecting Integration
Click the All Integrations tab in the Integrations menu, then click the app card you want to integrate.
Entering Integration Information
Enter additional integration information on the Install New Integration page.
- Name: Integration name
- Description: Integration description
- Depending on the Integration Type, additional information input may be required.
- Entering names and descriptions according to your purpose improves Tool call accuracy.
Installing OAuth-supported Apps
For MCP servers that support OAuth authentication like Google and Atlassian, clicking the Connect & Install button will redirect you to the login/authentication page of the corresponding application.
When the login page opens, select the account you want to use in AI Hub.
After authentication, the MCP server installation is completed.
Installing API Key-supported Apps
For MCP servers that don’t support OAuth authentication like AWS MCP, you need to obtain and enter API Keys such as Access Key and Secret Key. All API Key information is stored encrypted.
Clicking the Install button completes the integration with the entered API Key.
Installing n8n Workflow
Select the Chat Trigger Node in your existing workflow and copy the Chat URL and authentication information.
Select the n8n Chat Trigger Node type on the Integration page, enter the information you copied earlier, and save.
Creating MCP Presets
MCP Preset is a template that configures AI to operate in a specific context.
Go to the MCP Presets tab and click the Create New Preset button. (If you’re creating a preset for the first time, you’ll automatically enter the preset creation page.)
- Enter basic preset information.
- MCP Preset Name: A name to distinguish the preset.
- Instruction: Instructions that are basically included when chatting using the preset. It can include the preset’s role, context, response style, etc.
- In MCP Servers & Tools, select the tools of MCP servers to include in this preset.
- Click the Save Changes button to save.
By bundling various MCP tools into one preset, you can perform workflows across multiple SaaS platforms.
Chatting in AI Hub
Now all settings are complete. Start chatting with AI using MCP Presets!
- Go to the Chat tab.
- Type
@and select the MCP Preset you just created from the context menu. - Enter questions or requests in the chat box. For example, if you’ve integrated Slack MCP, the following requests are possible:
"Please send a message saying 'test' to the helloworld channel." - AI processes requests in real-time through the configured Integration.
If you’re curious about more MCP server features, check out the MCPs documentation.
Congratulations! You now understand the basic usage of QueryPie AI Hub.
If you’re curious about detailed features provided by each MCP server, check out the Available MCP Servers documentation.
Having Issues?
If you encounter problems or need additional help, please contact aihub@querypie.com.
Or click the chat icon at the bottom right of the QueryPie AI Hub homepage and leave a message.
We’ll provide quick support.