Skip to content

Getting Started for AI Builders

This guide walks you through publishing your first AI model on WasiAI and setting up monetization. By the end, you'll have a live model earning revenue from pay-per-inference and license sales.

Prerequisites

Before you begin, ensure you have:

  1. A wallet with some AVAX for gas fees (testnet AVAX for Fuji)
  2. Your AI model endpoint ready to deploy
  3. Model metadata prepared (name, description, category)
  4. Cover image (recommended: 1200x630px, PNG or JPG)
  5. Payment wallet address to receive micropayments

Tip: You can connect with Core Wallet, MetaMask, WalletConnect, or use social login (Google, Apple, Email) via Thirdweb.

Step 1: Connect Your Wallet

  1. Go to wasiai.io
  2. Click Connect Wallet in the top right
  3. Choose your connection method:
  4. Social Login: Google, Apple, Email, or Passkey
  5. Wallet: Core Wallet, MetaMask, WalletConnect, Coinbase Wallet

Once connected, you'll see your address in the navbar.

Step 2: Start the Publish Wizard

  1. Click Publish Model in the navigation
  2. You'll enter the 5-step Publish Wizard

Step 1: Basic Information

Fill in the essential details:

Field Description Example
Name Display name for your model "Crypto Sentiment Analyzer"
Slug URL-friendly identifier (auto-generated) "crypto-sentiment"
Short Summary Brief description "Analyzes sentiment of crypto news..."
Cover Image Visual representation Upload PNG/JPG
Business Category Industry category "Finance"
Model Type Type of AI model "Sentiment Analysis"

Step 2: Customer & Technical Details

Describe your model for users:

Field Description
Value Proposition What does this model do? (one sentence)
Industries Target industries
Use Cases Specific use cases
Inputs/Outputs What the model accepts and returns
Known Limitations Risks and limitations

Optional technical specs: Tasks, Modalities, Frameworks, Architecture.

Step 3: Artifacts & Inference

Upload files and configure your endpoint:

Field Required Description
Artifact Upload Yes At least one file with "primary-weights" role
Inference Endpoint Yes Your model's API URL
Payment Wallet Yes Address to receive USDC micropayments

Supported files: .bin, .pt, .onnx, .safetensors, .gguf, .zip, .json, .yaml

Step 4: Pricing

Set your monetization strategy:

Pricing Type Description Recommendation
Per-Inference Price per API call $0.001 - $0.10
Perpetual License One-time purchase $10 - $1,000

Royalties: Set a percentage (0-20%) you'll receive when the model is resold.

Step 5: Review & Publish

  1. Review all information
  2. Click Publish to Blockchain
  3. Confirm the transaction in your wallet
  4. Wait for confirmation (~1-2 seconds on Avalanche)

Step 3: Verify Your Listing

After publishing:

  1. Your model appears in the catalog
  2. An Agent ID is automatically created (ERC-8004)
  3. A Splitter contract is deployed for revenue distribution
  4. Users can now discover and use your model

Visit your model's detail page to see: - Live status - Agent ID badge - Pricing information - "Run Model" button

Monetization Options

Pay-Per-Inference (x402)

Users pay a small fee for each inference call. This is ideal for: - High-volume, low-cost models - Casual users who don't need unlimited access - Testing before committing to a license

How it works: 1. User clicks "Run Model" 2. Signs USDC authorization (gasless) 3. Payment is verified on-chain 4. Inference executes 5. You receive payment (minus fees)

License NFTs

Users purchase perpetual access. This is ideal for: - Power users who need unlimited access - Enterprise customers - Building recurring revenue

License type: - Perpetual: One-time payment, lifetime access

Revenue Distribution

When you receive a payment, it's automatically split:

Payment ($1.00)
    │
    ├── Marketplace Fee (5%): $0.05 → WasiAI
    ├── Creator Royalty (10%): $0.10 → Original Creator
    └── Seller Share (85%): $0.85 → You

Note: If you're both the creator and seller, you receive both shares.

Tracking Your Earnings

Dashboard

Access your creator dashboard to see: - Total revenue - Inference count - License sales - Reputation score

On-Chain Verification

All payments are verifiable on-chain: - View transactions on Snowtrace - Check your splitter contract balance - Verify license NFT mints

Updating Your Model

To publish a new version:

  1. Go to your model's detail page
  2. Click Publish Update
  3. Follow the wizard with updated information
  4. The new version joins your model family

Benefits of versioning: - Reputation carries over - Same splitter contract - Users can see version history

Best Practices

Pricing Strategy

Model Type Suggested Per-Inference Suggested License
Simple Classification $0.001 - $0.005 $10 - $50
Complex Analysis $0.01 - $0.05 $50 - $200
Generation (Text/Image) $0.05 - $0.20 $100 - $500

Metadata Quality

  • Write clear, specific descriptions
  • Use relevant keywords for discoverability
  • Include example inputs/outputs
  • Add documentation links

Reputation Building

  • Respond to user feedback
  • Maintain model accuracy
  • Update regularly
  • Engage with the community

Troubleshooting

"Transaction Failed"

  • Ensure you have enough AVAX for gas
  • Check that all required fields are filled
  • Try refreshing and reconnecting your wallet

"Model Not Appearing"

  • Wait 1-2 minutes for indexing
  • Refresh the catalog page
  • Check the transaction on Snowtrace

"Inference Not Working"

  • Verify your endpoint is accessible
  • Check the endpoint URL format
  • Test your endpoint directly with curl

Next Steps