How to create an Agent
Creating an Agent on AIsuru is your first step toward having a personalized virtual assistant. This process lets you define your Agent's core characteristics, from its identity to its AI capabilities.
Starting the creation process
To get started:
Go to AIsuru's main page;
Find and click the "Create" button in the "Your Agents" section.
Configuring the Agent
Once the process starts, you'll see a panel divided into four main sections.
All of these settings, except for the language, can be changed after the Agent is created. Don't worry too much if you're unsure about some of your initial choices.
General section
Here you'll define the identity and basic characteristics of your Agent:
Name: choose a representative name
Description: provide a short introduction to the Agent
Visibility:
Public: accessible to everyone, no password required
Private: requires a password
Language: choose the primary language
Voice type: choose between a male or female voice.
Categories: you can label your Agent by topic or purpose.
NSFW : enable only if your Agent handles adult content. Once activated, it can not be removed . Contact us if it was set by mistake.
Contextual Response: make your agent adapt its answers based on context:
Location: answers can vary depending on where the user is.
Date and time: answers can vary depending on the current day or time.
Accessible only to authenticated users: only logged-in users will get responses.
Advanced Settings
Ask for feedback on content-based responses: your agent will add an automatic prompt i.e "Was this helpful?" to collect feedback and improve content matching.
Disable AI Content Collection: prevents the agent from accumulating drafts and unanswered questions.
Suggest questions if the user is inactive: the agent proactively suggests questions when a user goes quiet.
Allow users to report incorrect responses: users can flag mismatched answers.
Confidence Threshold: controls how confident the agent must be before answering. This ensures more accurate answers but increases the number of times the agent says it doesn't know the answer.
Badges: DCM Integration: connect your agent to DCM badge integrations
Note: if you select "Public" during creation, the Agent will automatically appear on the homepage. You can remove it from the homepage by disabling the relevant option in the "Share me" section of your Agent after it's created.
AI section
This section lets you customize the AI's capabilities:
Response Mechanism:
Contents first, then AI (NLP + GenAI): the Agent looks for a match in your content first, then falls back to AI if nothing is found.
AI only (GenAI): the Agent always uses AI to generate responses.
Contents only (NLP): the Agent only replies using your uploaded content, with no AI
Choose Model: select which AI model powers your Agent.
Instructions : write a custom prompt to guide how your Agent behaves and responds. Use it to give your Agent a personality, a role or specific context
Deep Thought: Enable this feature for improved contextual understanding.
Deep Thought Instructions: Inform the AI what kind of information it should keep track of during the conversation
Advanced settings:
Configuration for import/export
Deep Thinking configuration (only available if the Deep Thinking feature is active)
Maximum number of tokens
Temperature.
Changing these parameters can significantly affect the quality of responses.
Appearance section
In this section you can customize the visual look of your Agent by uploading a cover image and an avatar. To do so:
Go to the "Appearance" section of the Agent creation panel;
For the cover image:
Click "Upload new cover" and choose the image you want to upload;
For the avatar:
Click "Upload avatar" and select your image;
After uploading the images, make sure to save your changes.
Supported formats are .png, .jpg, .jpeg, .gif with a maximum file size of 2 MB.
Extensions section
In this section you can choose which extensions to enable for the Agent that you're editing. But not to worry you can modify these settings later in the Functions and MCP sections.
Fetch URL: allows the Agent to visit web pages and extract relevant information from them.
Artifacts: the agents can create interactive and multimedia content during conversations.
Data Analysis: the agent is able to analyze data and generate insights.
Perplexity Search: uses AI to provide accurate and relevant search results.
Image Generator: allows the Agent to generate images.
Send Email: allows the Agent to send emails automatically.
Finishing up
Once you've configured all the sections:
Review the settings you've chosen
Click "Create agent" to create your Agent.
Creating an Agent is the first step toward a personalized virtual assistance experience. Experiment with different configurations to find the one that best fits your needs. 🚀
Last updated