Axiom Stack

Payment Factory use cases

Six customer shapes Payment Factory ships for today — one dogfooded by Themra, five pattern-templates ready for any Solana-anchored service.

01Live · dogfood

Oracle data

Our own dogfood. Charge agents per attestation.

Themra uses Payment Factory to charge per attestation. Live x402 manifest at api.axiomstack.dev/v1/.well-known/x402 — try it; it is the same surface every other service gets.

curl -X POST https://api.axiomstack.dev/v1/attestation/instant \
  -H "Content-Type: application/json" \
  -d '{"asset_class": 6, "asset_id": "BTC"}'

# HTTP 402 → Payment-Required envelope with x402 / MPP / AP2 / Bearer options
# Re-submit with chosen auth method → HTTP 200 with attestation + on-chain PDA

Pricing guidance

Equity & crypto instant: $0.02 per call (x402) · audit tier: $0.10. Bearer + monthly quota for enterprise consumers.

02Pattern · bring your own service

API access

Charge per API call. Your service publishes a manifest; agents pay before invoking.

Wrap any HTTP endpoint behind Payment Factory. Publish a .well-known/x402 manifest at your domain advertising your payTo (treasury owner pubkey), price, and accepted protocols. No subscription gate, no commitment — agents pay per call.

// Your service's /.well-known/x402 manifest
{
  "x402Version": 1,
  "accepts": [{
    "scheme": "exact",
    "network": "solana-devnet",
    "asset": "<USDC mint>",
    "maxAmountRequired": "5000",
    "payTo": "<YOUR_TREASURY_OWNER_PUBKEY>",
    "resource": "https://yourapi.dev/v1/endpoint"
  }]
}

// On HTTP 402, return the same payload + Payment-Required header.
// Settlement hits YOUR treasury — Axiom backend is uninvolved.

Pricing guidance

You set the price per call. Suggested floor: $0.001 (covers Solana TX cost). Typical: $0.01–$0.10 per call.

03Pattern · bring your own service

Content licensing

Charge per image, dataset, or article fetch.

Per-resource pricing via .well-known/x402 manifest — each licensed resource gets its own price. Useful for image archives, structured datasets, syndicated content, premium feeds. Agents pay per fetch; you get on-chain settlement + a receipt UI.

// Resource-scoped pricing in the manifest
{
  "accepts": [
    {
      "resource": "https://yoursite.dev/api/dataset/sample",
      "maxAmountRequired": "1000",
      "asset": "<USDC mint>",
      "payTo": "<your-treasury>"
    },
    {
      "resource": "https://yoursite.dev/api/dataset/full",
      "maxAmountRequired": "500000",
      "asset": "<USDC mint>",
      "payTo": "<your-treasury>"
    }
  ]
}

Pricing guidance

Per-resource pricing — e.g., $0.001 sample, $0.50 full dataset. Bearer + bundle pricing layer at V1.1.

04Pattern · bring your own service

Compute · LLM inference, GPU rentals

Charge per inference token or compute-minute.

For session-style billing (multi-turn LLM, streaming GPU compute), MPP-charge intent lets a customer agent pre-authorize a budget and your service settles per unit. AP2 mandates let an upstream operator pre-delegate a spending cap. Both reduce per-call friction for high-throughput compute.

// Customer agent acquires an MPP-charge intent
const intent = await client.sign_mpp_charge({
  challenge_id: "session-2026-05-27-001",
  intent: "compute:gpt-4-equivalent:1000-tokens",
  amount_micros: 250000, // $0.25 budget
  to: "<your-treasury-owner>"
});

// Your service settles draws against the intent as inference proceeds
// Receipts written to oracle-mpp-receipts; auditable in dashboard

Pricing guidance

MPP-charge intent per session ($0.10–$5 typical pre-auth); pay-per-minute or pay-per-token settlement.

05Pattern · bring your own service

Model inference

Sell access to your fine-tuned model per inference call.

For proprietary or fine-tuned models — your model lives behind a paywall, agents pay per generation. x402 for per-call pricing; AP2 mandate for delegated budgets where an upstream operator authorizes the agent to spend up to X on inferences.

# Python (axiom-stack SDK 0.5.0+)
from axiom_stack import Client

client = Client(api_key=None)  # x402-only flow, no Bearer

result = client.pay(
    to="<your-treasury-owner>",
    amount_micros=20000,            # $0.02 per inference
    protocol="x402",
    network="solana-devnet",
    asset="<USDC mint>",
    resource="https://yourmodel.dev/v1/generate"
)
# After settlement, re-POST your inference request with PAYMENT-SIGNATURE

Pricing guidance

Per-inference call: $0.001–$1.00 depending on model size. Premium fine-tuned models: AP2 mandate for budgeted consumption.

06Pattern · bring your own service

Agent-to-agent commerce

Agents pay each other for delegated tasks.

Multi-hop settlement: an upstream agent delegates a task to a downstream agent, which may further delegate to specialized agents. Each hop settles on-chain via x402 or against an AP2 mandate the upstream operator signed. Receipts trace the full delegation chain.

// Upstream agent calls a downstream agent's task endpoint
// Downstream agent's manifest:
{
  "accepts": [{
    "scheme": "exact",
    "network": "solana-devnet",
    "payTo": "<downstream-agent-treasury>",
    "maxAmountRequired": "100000",
    "asset": "<USDC mint>",
    "resource": "https://downstream-agent.dev/api/task"
  }]
}

// Upstream agent uses the SDK to pay + invoke
const result = await client.pay({
  to: "<downstream-agent-treasury>",
  amount_micros: 100_000,
  protocol: "x402",
  resource: "https://downstream-agent.dev/api/task"
});

Pricing guidance

Per-task pricing set by each agent. Typical: $0.01–$10 per delegated task. Receipts trace delegation chains.