Token Factory: Launch AI-Powered Tokens on Base
Deploy tokens that are backed by real AI compute from day one. Your token isn't just a ticker—it's a key to the AIPG network.
The Problem
Launching a token today means:
- Deploying a generic ERC-20 with no utility
- Building your own infrastructure from scratch
- Convincing users your token does something
- No connection to real services or compute
Most tokens ship with a promise and a roadmap. Yours ships with working AI infrastructure.
What You Get
When you deploy through AIPG's Token Factory, your token is wired into a production AI network from block one:
| Capability | Description |
|---|---|
| Grid Access | Your users can pay for AI compute with your token |
| Model Registry | Access 50+ registered AI models (Flux, SDXL, LLMs) |
| Recipe Library | Thousands of ComfyUI workflows ready to use |
| Worker Network | Decentralized GPU workers fulfill jobs |
| Job Anchoring | On-chain Merkle proofs for every generation |
| Revenue Share | Earn AIPG when your token is used for compute |
Your token becomes a compute credit for real AI services.
How It Works
1. Deploy Your Token
Choose a template and configure:
Token Name: MyAIToken
Symbol: MAIT
Supply: 100,000,000
Compute Ratio: 1 MAIT = 0.1 AIPG worth of compute
Revenue Split: 80% to your treasury, 20% to AIPG networkThe factory deploys your token with built-in hooks to the AIPG Grid.
2. Users Spend Your Token
When someone uses your token for AI generation:
┌─────────────────────────────────────────────────────────────┐
│ User pays 10 MAIT for image generation │
├─────────────────────────────────────────────────────────────┤
│ → MAIT is burned or sent to your treasury (configurable) │
│ → Grid receives job request with payment proof │
│ → Worker generates image using registered model │
│ → Result anchored on-chain with Merkle proof │
│ → User receives deterministic, verifiable output │
└─────────────────────────────────────────────────────────────┘3. Revenue Flows Back
Every job creates value:
- Your Treasury: Receives the MAIT spent (or burn reduces supply)
- GPU Workers: Earn AIPG for compute work
- AIPG Network: Protocol fee sustains infrastructure
- Your Holders: Token has real utility = real demand
Token Templates
Compute Credit Token
For projects that want a dedicated AI compute currency.
- Users buy your token to access AI features
- Token is burned on use (deflationary)
- Price discovery based on compute demand
- Example: An AI art platform with its own economy
Access Token
For projects that want subscription-style access.
- Holding X tokens unlocks AI features
- No burning—just holding requirement
- Tiered access based on balance
- Example: AI tool suite with membership tiers
Hybrid Token
For projects that want staking + compute.
- Stake tokens to earn compute credits
- Unstaked tokens can pay for compute directly
- Dual utility: passive yield + active use
- Example: AI agent that rewards long-term holders
Agent Token
For autonomous AI agents with on-chain treasuries.
- Agent controls its own wallet
- Earns revenue from services
- Spends on compute to fulfill tasks
- Example: Trading bot, content generator, research agent
Integration Architecture
Your token plugs into AIPG's existing production contracts:
┌────────────────────────────────────────────────────────────────────┐
│ YOUR TOKEN │
│ (ERC-20 deployed via Token Factory) │
└─────────────────────────────┬──────────────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────────────────┐
│ PAYMENT ROUTER │
│ Converts your token → compute credits │
│ Handles burns, treasury splits, revenue share │
└─────────────────────────────┬──────────────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────────────────┐
│ AIPG GRID (Diamond Proxy) │
│ 0x79F39f2a0eA476f53994812e6a8f3C8CFe08c609 │
├────────────────────────────────────────────────────────────────────┤
│ │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ ┌────────────┐ │
│ │ ModelVault │ │ RecipeVault │ │ JobAnchor │ │ Worker │ │
│ │ │ │ │ │ │ │ Registry │ │
│ │ 50+ models │ │ Workflows │ │ Merkle │ │ │ │
│ │ registered │ │ stored │ │ proofs │ │ GPU nodes │ │
│ └─────────────┘ └─────────────┘ └─────────────┘ └────────────┘ │
│ │
└────────────────────────────────────────────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────────────────┐
│ GPU WORKER NETWORK │
│ Decentralized nodes running Flux, SDXL, LLMs │
│ Bonded with AIPG stake, slashable for misbehavior │
└────────────────────────────────────────────────────────────────────┘Technical Details
Token Contract
Factory-deployed tokens include:
interface IFactoryToken {
// Standard ERC-20
function transfer(address to, uint256 amount) external returns (bool);
function approve(address spender, uint256 amount) external returns (bool);
// Compute integration
function computeBalance(address user) external view returns (uint256);
function spendForCompute(uint256 amount, bytes32 jobId) external;
// Revenue tracking
function totalComputeSpent() external view returns (uint256);
function totalRevenue() external view returns (uint256);
// Factory metadata
function computeRatio() external view returns (uint256);
function revenueRecipient() external view returns (address);
}Payment Flow
When a user spends your token for AI compute:
- Authorization: User approves PaymentRouter for their tokens
- Job Submission: User calls
submitJob(modelId, recipeId, params, paymentToken, amount) - Payment Processing: Router converts token amount to compute units
- Token Handling: Based on config—burned, sent to treasury, or split
- Job Execution: Grid assigns job to bonded worker
- Completion: Worker submits result, anchored via Merkle root
- Settlement: Worker earns AIPG, your treasury earns revenue share
Supported Models
Your token holders get access to all registered models:
| Category | Examples | Use Case |
|---|---|---|
| Image Gen | Flux.1, SDXL, SD 1.5 | Text-to-image, img2img |
| ControlNet | Canny, Depth, Pose | Guided generation |
| LoRA | Style adapters, character models | Fine-tuned outputs |
| Inpainting | Flux Inpaint, SDXL Inpaint | Edit regions |
| Upscale | 4x-UltraSharp, ESRGAN | Resolution enhancement |
| LLM | Llama, Mistral, Qwen | Text generation, chat |
Models are registered in ModelVault with full metadata: VRAM requirements, supported features, constraints.
Workflows
Recipes stored in RecipeVault define complete ComfyUI pipelines:
- Multi-model compositions
- ControlNet + LoRA stacking
- Post-processing chains
- Deterministic reproduction via stored parameters
Your users can run any public recipe or create private ones.
Economics
For Token Creators
| Benefit | Details |
|---|---|
| Instant Utility | Token works from block 1 |
| Revenue Stream | Earn on every compute job |
| No Infra Costs | AIPG network handles compute |
| Deflationary Option | Burn on use if desired |
| Network Effects | More users = more demand |
For Token Holders
| Benefit | Details |
|---|---|
| Real Use Case | Pay for AI with your bag |
| Price Support | Utility creates buy pressure |
| Access to AI | No API keys, no subscriptions |
| Verifiable Output | On-chain proofs for every job |
Revenue Split Example
User pays 100 MYTOKEN for a Flux generation:
┌─────────────────────────────────────────┐
│ 100 MYTOKEN paid │
├─────────────────────────────────────────┤
│ → 80 MYTOKEN to your treasury │
│ → 10 MYTOKEN burned (supply reduction) │
│ → 10 MYTOKEN protocol fee (buys AIPG) │
├─────────────────────────────────────────┤
│ Meanwhile: │
│ → GPU worker earns 0.5 AIPG for job │
│ → Job anchored on-chain forever │
│ → User gets deterministic image │
└─────────────────────────────────────────┘Deployment Flow
Step 1: Configure
# Via CLI or web interface
aipg-factory create \
--name "MyAIToken" \
--symbol "MAIT" \
--supply 100000000 \
--compute-ratio 0.1 \
--burn-percent 10 \
--treasury 0xYourAddress \
--revenue-split 80Step 2: Deploy
Factory deploys:
- Your ERC-20 token contract
- Payment routing configuration
- Revenue split settings
- Grid access permissions
Step 3: Integrate
import { AIPGFactory } from '@aipowergrid/sdk';
// Initialize your token
const myToken = await AIPGFactory.connect(tokenAddress);
// User generates an image
await myToken.generateImage({
model: 'flux-schnell',
prompt: 'a robot painting a sunset',
paymentAmount: 10, // 10 MYTOKEN
});
// Check job status
const job = await myToken.getJob(jobId);
console.log(job.status, job.imageUrl, job.merkleProof);Step 4: Launch
- Distribute tokens to users
- They can immediately use for AI compute
- Revenue flows to your treasury
- Monitor usage via Grid events
Use Cases
AI Art Platform
Launch a token for your generative art community:
- Artists pay with your token to create
- NFT minting with on-chain provenance
- Deflationary tokenomics via burn-on-mint
- Revenue funds community grants
AI Agent Network
Deploy tokens for autonomous agents:
- Each agent has its own token treasury
- Agents earn by providing services
- Spend tokens on compute to fulfill tasks
- Transparent on-chain economics
AI Tool Suite
Subscription access to AI tools:
- Hold tokens for tiered access
- No per-use fees for holders
- Staking unlocks premium features
- Token buybacks from revenue
Game Economy
In-game currency with AI generation:
- Generate game assets with tokens
- Verifiable scarcity via on-chain anchoring
- Player-owned economy
- AI NPCs powered by compute spending
Security
Inherited from AIPG
Your token benefits from AIPG's production infrastructure:
- Audited Contracts: Core Grid contracts reviewed
- Access Control: Role-based permissions throughout
- Reentrancy Guards: All state changes protected
- Pausable: Emergency stops on all contracts
- Bonded Workers: GPU nodes stake AIPG, slashable
Factory-Specific
- Immutable Config: Core parameters locked at deploy
- Revenue Routing: Trustless splits via smart contracts
- Rate Limits: Configurable spam protection
- Allowlists: Optional restricted access
Roadmap
| Phase | Status | Features |
|---|---|---|
| Contracts | In Development | Token factory, payment router |
| CLI Tool | Planned | aipg-factory deployment tool |
| Web UI | Planned | No-code token deployment |
| SDK | Planned | JavaScript/Python integration |
| Templates | Planned | Pre-built token configurations |
Get Started
The Token Factory is under active development. To get early access or discuss your project:
- Discord: Join the community (opens in a new tab)
- GitHub: github.com/aipowergrid (opens in a new tab)
- Twitter: @aipowergrid (opens in a new tab)
FAQ
Q: Do I need AIPG to launch a token? A: No. You deploy your own token. AIPG is used internally for worker rewards, but your users only interact with your token.
Q: What chains are supported? A: Base (mainnet) initially. Other L2s may be added based on demand.
Q: Can I migrate an existing token? A: Yes. We can add compute hooks to existing ERC-20s via wrapper contracts.
Q: What's the cost to deploy? A: Gas costs on Base (~$1-5) plus a one-time factory fee (TBD, likely small AIPG amount).
Q: How do workers get paid? A: Workers always receive AIPG. The payment router handles conversion from your token.
Q: Can I run my own workers? A: Yes. Register your GPUs in WorkerRegistry, bond AIPG, and earn from jobs paid in any factory token.