How to bring your own Deepseek model into Momen (BYOM)?
Introduction
Large Language Models (LLMs) are evolving rapidly, but their true potential lies in how they are applied. Custom AI applications tailored to your needs can set your business apart in this AI-driven era.
AI is becoming more accessible than ever. Even if you’re not a developer, you can still leverage its power. Momen is a full-stack, no-code web app builder that empowers non-technical founders to create AI applications. With Momen’s AI agent builder, you can design AI apps and intelligent agents customized for your business. Now, with the launch of BYOM (Bring Your Own Model), you can integrate different AI models into your apps and workflows.
What is BYOM?
BYOM stands for Bring Your Own Model, allows you to deploy your own AI models on cloud platforms and integrate them into your applications. In Momen, this process is simplified—just enter your cloud provider’s endpoint details, and you’re ready to integrate your model into your AI workflows.
This guide walks you through deploying DeepSeek R1 on Azure AI Foundry and integrating it into Momen using BYOM.
Why Use BYOM?
Different AI models excel in different areas. DeepSeek R1 is a cost-effective model with strong reasoning capabilities, but it lacks multimodal support and tool invocation (as of March 2025). On the other hand, GPT-4o is better suited for tasks requiring multimodal capabilities and tool invocation, such as image processing and agentic workflow.
By using Momen’s BYOM feature, you can deploy DeepSeek R1 while also integrating GPT-4o—combining their strengths to build powerful AI applications. This is exactly how we built "Angry Dietitian," an AI-powered diet assistant that leverages DeepSeek for reasoning and GPT-4o for image processing.
How to Integrate DeepSeek into Momen
Follow the steps below to get started.
Step 1: Create an Azure Account & Subscription
Go to Azure’s official website and sign up for an account.
After logging in, create a subscription if you don’t already have one.
Step 2: Create an AI Hub in Azure AI Foundry
Search for Azure AI Foundry in the Azure portal or go directly to it.
Click "Create a Hub" to start the setup.
Step 3: Configure Your AI Hub
Fill in the required information for the AI Hub setup.
If you haven’t created a Resource Group or AI Services before, select "Create New" to generate them automatically.
Most default settings work for basic deployment, so unless you have specific requirements, leave them unchanged.
Click "Review + Create", then wait a few minutes for the deployment to complete.
Step 4: Access AI Hub and Create a Project
Once the AI Hub is created, navigate to it.
Click "Launch" to enter Azure AI Foundry.
Inside Foundry, create a new project for DeepSeek R1.
Step 5: Deploy DeepSeek R1
In the project, go to Models + Endpoints (bottom left corner).
Select "Deploy a Base Model", choose DeepSeek R1, and confirm the deployment.
Wait for the deployment to complete.
Step 6: Integrate DeepSeek into Momen
Once the model is deployed, retrieve its endpoint and API key from Azure. Then, go to Momen’s AI Agent Builder and add DeepSeek R1 as a custom model using the retrieved information.
Now, you’re ready to use DeepSeek within Momen!
Note: When using BYOM (Bring Your Own Model) in Momen, you are responsible for covering the token costs of your deployed model. Unlike the default GPT integration, which allows users to top up AI points within Momen, models like DeepSeek R1 deployed via Azure will be billed separately based on usage. Be sure to check your cloud provider's pricing to manage costs effectively.
Conclusion
With Momen’s BYOM feature, you can seamlessly integrate your own AI models into agentic workflows—without writing a single line of code. Whether you're technical or not, you can leverage different AI models and build them into powerful business applications.
Check out how we built "Angry Dietitian" using DeepSeek R1 for reasoning and GPT-4o for image processing:
