Quickstart Guide
Everything you need to know to get started creating AI Agents.
Last updated
Everything you need to know to get started creating AI Agents.
Last updated
In this article, we’ll be guiding you through everything you need to know to get started on MindStudio and teaching you to create your first AI Agent.
After creating an account, you'll land on the Workspace Overview Screen, where you can view all published AI Agents. If there are no AI Agents published, then you will see the Getting Started Guide.
From the Workspace Overview Screen, you’ll find several controls on the left. From here you can:
Create new workspaces.
In this guide, we’ll build an example AI Agent to find daily tech news, summarize it, and sends an email every morning at 8:00 AM.
To create an AI Agent, click on the Create New Agent button at the top-right. This will create a new Agent and open the MindStudio Editor.
The editor is made of two key areas: the Explorer Tab and the Navigator.
Explorer Tab
This is where you'll find all of the resources used to build your AI Agents. This includes resources like Data Sources, Functions, User Inputs, and Workflows
Navigator (Main Workspace)
The large area covering the rest of the Editor if the Main Workspace, also refer to as the Navigator. This area will change depending on what you have selected. By default, the Editor will open on the Main.flow workflow and have the Prompt Tab open.
The System Prompt serves as the AI Agent's core instructions, defining its role, capabilities, and constraints.
You can write a system prompt manually by typing into the blank space below, or you can click on the Generate Prompt button at the bottom left to have the Prompt Generator to write the prompt for you.
Click on the Generate Prompt button at the bottom left of the prompt area. This will open the Prompt Generator Modal.
Using natural language, enter a brief description of what your AI Agent is supposed to do. For this build, we can type something like:
After entering your description, click the Generate button. The system will automatically create a structured prompt based on your input.
Carefully review the generated prompt to make sure it aligns with your requirements. You can edit the prompt directly in the editor if adjustments are needed.
Once your System Prompt is written, navigate to the Automations tab.
The Automations Tab is where you design the workflow for your AI Agent. This section allows you to define the sequence of actions your AI Agent will follow.
Start Block:
Schedule the workflow to run at 8:00 AM daily.
Define a launch variable called topic.
Google Search Block:
Use a function to perform a Google search.
Reference the topic variable as the search query.
Save the result as a new variable, google_result.
Text Generation Block:
Prompt AI to summarize the google_result.
Use the prompt:
Terminator End Block w/ Email Notification:
Enable Email notifications to send the summary to the specified email.
MindStudio provides access to over 50 AI models from leading providers, allowing you to tailor your workflows for a variety of tasks and use cases.
Access Model Settings
Navigate to the Model Settings Tab in your workflow editor.
Choose a Model
Select from over 50 AI models available.
NOTE: Some models may be locked and can be unlocked by adding a payment method in your Billing Settings.
There are two configuration options for AI models:
Temperature
Controls the randomness of responses:
Higher Temperature: More creative, but less predictable.
Lower Temperature: More consistent, but less diverse.
Max Response Size
Defines the maximum size of the model’s response in tokens.
Example: 400 tokens ≈ 300 words.
Every block that uses AI has the ability to override the underlying models settings and use its own unique AI model and configuration. This allows you to use different models in the workflow based on the task required at each specific step.
Testing and evaluation are critical steps to ensure your AI Agent performs as expected. MindStudio provides tools to debug, validate, and optimize your AI Agents.
Navigate to the Errors Tab to view any issues, such as misspelled variables or misconfigured blocks.
Click on an error in the Errors Tab to highlight the problematic block in the workflow.
Adjust configurations or correct spelling to resolve the issue.
Once fixed, errors will disappear from the tab.
Open the Evaluations Tab to create test scenarios for your workflow.
Click the Autogenerate button to generate test cases automatically. You can provide additional context for test case generation, such as specifying genres, topics, or scenarios.
Specify how many test cases you want to generate.
Review the generated cases to ensure they cover a variety of inputs
Use the Run button in the top-left to execute the workflow multiple times with the test cases.
After execution, review the results in the right-hand column. Check whether the outputs align with expected behavior.
Once your AI Agent has been tested and evaluated to ensure it functions correctly, the next step is to publish it. Publishing makes your AI Agent available for use and integration.
Set Metadata
Navigate to the Explorer Tab and click on the root menu.
Add the following details:
Name: Give your AI Agent a descriptive and unique name.
Description: Provide a short explanation of what your AI Agent does.
Go through other sections and configure other settings.
Publish the Workflow
Click the Publish button to finalize your AI Agent.
Once published, the AI Agent will be accessible via the Published Tab in your workspace. Click on the confirmation message to view your AI Agent.
Integration Options
After publishing, you can integrate your AI Agent with external tools and platforms, such as:
Zapier: Automate workflows across apps.
Make: Build advanced automation scenarios.
API: Use the provided documentation to connect via custom APIs.
Test Request Integration
Use the Create Test Request feature to validate API calls.
Configure the API request by:
Selecting a workflow.
Providing launch variables (e.g., topic = "daily tech news").
Run the test and review the logs for successful execution.
View the AI Agent in the Workspace Overview
View the newly published AI Agent under the Home Tab.
Confirm it is listed alongside other published Agents.
Congratulations on creating and publishing your first AI Agent in MindStudio! We’ve covered everything from writing system prompts and building workflows to testing, debugging, and publishing. MindStudio is a powerful and versatile platform, enabling you to create AI Agents tailored to specific needs.
Remember, this guide is just the beginning. The possibilities with MindStudio are endless, and we’re excited to see what you’ll build. If you ever need help, our Support Chat is just a click away.
Before starting, you will need to .
Access general .
Explore to view AI Agent activity.
Manage (e.g., budget limits, payment methods).
to your workspace.
Access like API Keys, Request Logs, and Documentation.
