Getting Started

Welcome to Tunaas.ai! This guide will help you navigate our platform and create your own AI models without any coding experience.

Creating Your Account

1. Click the Sign Up button in the top right corner
2. Fill in your username, email, and password
3. Verify your email address
4. Log in to access your dashboard

Setting Up Your Workspace

1. Customize your profile in the Account Settings
2. Connect your Hugging Face account (optional but recommended)
3. Explore the dashboard to understand available features

Creating Models

Creating a custom AI model starts with selecting the right base model and defining its purpose.

Step-by-Step Guide

1. From your dashboard, click Create New Model
2. Choose a name and add a description for your model
3. Select a base model type:
   • Text Generation: For creating conversational agents, content creators, etc.
   • Text Classification: For categorizing text into predefined classes
   • Summarization: For condensing long texts while preserving key information
   • Question Answering: For AI that responds to specific queries
4. Choose a specific base model - we recommend starting with smaller models if you're new
5. Click Create Model to proceed to the fine-tuning stage

Fine-Tuning Your Model

Fine-tuning is the process of training an existing model on your specific examples to customize its behavior.

Adding Training Examples

1. Navigate to your model's detail page
2. Click Add Training Examples
3. For text generation models:
   • In the Input Text field, enter a prompt or question
   • In the Output Text field, enter the ideal response
4. For classification models:
   • Enter text examples and assign them to your predefined categories
5. Add at least 5 diverse examples for best results
6. Click Save Example after each entry

Batch Upload

For larger datasets, use the Batch Upload feature:

1. Prepare your examples in JSON format
2. Click Upload Training Data
3. Paste or upload your JSON file
4. Verify your data was imported correctly

Starting the Fine-Tuning Process

1. Once you've added at least 5 examples, click Start Fine-Tuning
2. Select your subscription plan if prompted
3. Review the fine-tuning configuration
4. Click Begin Fine-Tuning to start the process
5. The process can take 15-60 minutes depending on model size and example count

Using Datasets

Instead of creating training examples manually, you can import existing datasets from Hugging Face.

Finding and Importing Datasets

1. Go to the Datasets tab in the navigation menu
2. Use the search function to find relevant datasets for your task
3. Click on a dataset to view its details
4. Click Import Dataset to add it to your account
5. After importing, you can use it for any of your models

Using a Dataset for Fine-Tuning

1. From your model's detail page, click Use Dataset
2. Select an imported dataset from the dropdown
3. Configure how the dataset should be used (e.g., selecting specific columns)
4. Click Apply Dataset to use it for fine-tuning

Prompt Templates

Prompt templates help structure inputs to your model in a consistent way to get better outputs.

Creating Prompt Templates

1. Go to Prompt Templates in the navigation menu
2. Click Create New Template
3. Choose a template type or use our AI-generated templates
4. Define variables in your template using {variable_name} syntax
5. Save your template

Using AI-Generated Templates

1. Click Generate with AI in the Prompt Templates section
2. Select the type of template you need (coaching, teaching, etc.)
3. Add specific context if needed
4. Click Generate Template
5. Review, customize, and save the generated template

Applying Templates to Deployments

1. When creating a deployment, select a prompt template
2. The template will be used to format inputs before they're sent to your model
3. You can change templates for existing deployments at any time

Deployments

After fine-tuning, deploy your model to make it accessible for real-world use.

Deployment Options

1. From your fine-tuned model page, click Deploy Model
2. Choose a deployment platform:
   • Hugging Face Inference API: Cloud-hosted with API access
   • Ollama: For local deployment on your own hardware
   • Model Context Protocol (MCP): For creating context-aware AI agents that integrate with your applications
   • Replit App: Hosted web application with custom domain
   • Vercel Serverless: Serverless deployment with high scalability
3. Select an interface type:
   • API Endpoint: For integration with your own applications
   • Chat Interface: Ready-to-use chat UI for conversations
   • AI Agent: Advanced interface with tool-using capabilities
   • Embeddable Widget: For embedding on websites with minimal code
   • MCP Server: Context-aware AI server that implements the Model Context Protocol
   • Local Deployment: For running on your own hardware with Ollama
4. Configure deployment-specific settings
5. Click Deploy to create your deployment

Managing Deployments

From the Deployments page, you can:

• View usage statistics and logs
• Update configuration settings
• Pause or delete deployments
• Get access credentials and integration instructions

Marketplace

The Tunaas.ai Marketplace allows you to buy and sell fine-tuned models.

Browsing the Marketplace

1. Go to the Marketplace tab in the navigation
2. Use filters to narrow down by category, price, or rating
3. Click on models to view detailed descriptions, reviews, and examples
4. Purchase models directly through our secure payment system

Selling Your Models

1. Create and fine-tune a model as usual
2. From your model's page, click List on Marketplace
3. Set a price and add detailed marketplace information
4. Submit for review (usually completed within 24 hours)
5. Once approved, your model will be available for purchase
6. You receive 80% of each sale

Subscription Plans

Tunaas.ai offers tiered subscription plans to fit different needs and budgets.

Frequently Asked Questions