Step-by-Step Guide: Lyzr Agents
Lyzr Agents make building your own safe & responsible agents very easy. It gives tough competition to platforms like Relevance AI, Mind Studio, Flowise, and LangFlow.
Lyzr Agents is building a platform that would help companies with a reliable way to build safe and responsible agentic workflows. It’s a no-code platform for building agents similar to Relevance AI, Mindstudio , Flowise, Langflow and many others out there. However they claim to be the one especially focussed on Reliable and Safe AI.
I have reviewed Mindstudio and RelevanceAI earlier and wanted to give this a shot. Please read those articles here:
Mindstudio:
Relevanceai:
Also read my post on the best AI Frameworks to work on
Coming to Lyzr Agents, they have a Lyzr Agent Studio where any one can build simple agents. You can check out the video here or check out the step-by step guide.
Step by Step guide Lyzr Agents
Step 1: Access the Agent Creation Interface
Navigate to the 'Agents' Section: In the Lyzr interface's left-hand navigation menu, click on the "Agents" option.
Click 'Build Agent': On the main dashboard area or on the 'Agents' page, find and click the "+ Build Agent" button. This opens the Agent creation form.
Step 2: Define Basic Agent Details
Agent Name: In the "Name" field, enter a descriptive name for your agent (e.g., "Study Abroad Newsletter Writer").
Agent Description: In the "Description" field, provide a brief explanation of what the agent does (e.g., "You are a study abroad research agent that creates a newsletter on what's happening in the world of Study Abroad").
Step 3: Select LLM Provider and Model
LLM Provider: In the "LLM Provider" dropdown, choose the desired provider (e.g., "OpenAI").
LLM Model: In the "LLM Model" dropdown, select the appropriate model (e.g., "gpt-4o-mini").
Step 4: Define Agent Role and Instructions
Agent Role: In the "Agent Role" field, describe the agent's primary function or expertise (e.g., "You are a Study Abroad Newsletter Writer").
Agent Instructions: In the "Agent Instructions" text area, provide clear, concise instructions.
Step 5: Improve Instructions
Click Improve In the "Agent Instructions" text area. This will automatically update the instructions by improving them for the agent.
Step 6: Add Tools and Configuration
Select a Tool: Use the "Select a tool" dropdown to choose the external tool you want the agent to use (e.g., "Perplexityai").
How would you use this tool? Explain how you want the agent to use this tool, including the specific purpose (e.g., "Would like to use this tool for research").
Step 7: Configure Core Features & Safe AI
Humanizer: Turn on the Humanizer Feature.
Fairness and Bias: Turn on the Fairness and Bias feature.
Step 8: View API Endpoint
Click Agent API In the top right of the screen, there is a menu item that allows you to view the Agent API as a JSON, Inference, or to create it.
Step 9: Create the Agent
Click Create: Locate and click the "Create" button. This will finalize the agent creation process.
Step 10: Test the Agent
Test Agent Inference Once created, use the "Test Agent Inference" panel to interact with the agent and verify its functionality. Enter a query and click the send button.
Step 11: Launch as App
Select Launch as App In the top right of the screen, select to Launch the Agent as App.
Select Yes I Am Ready The button may be behind a confirmation modal asking if you are satisfied with the agent's quality of the inference.
Enter App information Once the button is selected, a modal will appear that will ask for the app information such as App Name, a Description, Industry, Function, Visibility, and Category.
Launch App After all the appropriate information is selected, click the "Launch app" button and the app will be created successfully!
Step 12: Test the Agent and Share
Go to Agents Section Once created, check the agents section and see that the app has been created.
Share and test Share and test the application to verify functionality.
This guide should help you recreate the agent creation process shown in the video. Remember to check the Lyzr documentation for more detailed explanations of each setting and feature.
After exploring Lyzr Agents and reviewing demonstrations on their social media, I've found it to be a promising platform.
While the showcase videos hint at sophisticated, custom-built agents, replicating that level of complexity within the core Lyzr Agent Studio proved challenging, suggesting a difference between demo and typical user experience.
Check out few interesting Demo Videos
Key Strengths of Lyzr Agents
Simplicity and Ease of Use (For Basic Use Cases): The core platform interface is relatively straightforward, making it accessible for defining and configuring agents aimed at general tasks.
API Integration Potential: The availability of comprehensive APIs is a major plus. This allows developers to deeply integrate with Lyzr Agents, customize functionality beyond the basic interface, and quickly deploy agent-driven features.
Model Provider Flexibility: Lyzr Agents support connecting to a broad range of LLM providers, including OpenAI, Anthropic, Google, and Deepseek. This gives both individual users and enterprise teams the freedom to select the model that best aligns with their performance needs and cost considerations.
Agent Marketplace for Inspiration and Deployment: The Agent Marketplace is a standout feature. The marketplace showcases agents created by both the Lyzr team and other users. These agents serve as useful templates that can be deployed with a single click or serve to provide inspiration and learning experiences.
Areas of Improvement for Lyzr Agents
Limited Tool Selection: The range of available tools for agent integration is currently somewhat restricted, with roughly 15 options in total. Even for research tasks, the options are limited and may lack specific versatility. Essential tools like Gmail, Slack, and Notion are present, but a more extensive library would significantly expand agent capabilities and applicability.
Pricing Plan
You can create unlimited AI agents in all the plans which is wonderful. The pricing looks competitive.
Lyzr Agents is one platform to watch out for and has a great potential to create an impact in the evolving Agents Space. But I think it has some catchup to do especially when it comes to Tools Marketplace and improving the flexibility of the platform to build more advanced AI agents.
Website: https://www.lyzr.ai
Funding: 2.7Million USD