Payments and pricing
Developer-side payments
Billing is intent-based:
- Create a payment intent with a method the operator turned on (direct Monad, x402, or future card rails).
- Confirm with either:
- Direct Monad — send funds to the shown receive guidance; confirm with your transaction hash when the UI asks.
- x402 — the API may return payment required with structured headers; the portal completes the facilitator flow for you.
Clients must handle structured payment challenges—not every error is a plain JSON message.
Treasury addresses and networks come from deployment configuration for each environment.
Pricing model
Pricing is data-driven in the operator console, not hardcoded in the portal UI:
- Categories define the base catalog (also reflected on the public landing).
- Rules encode complexity and other multipliers.
- Campaigns can run first-N or windowed discounts with counters.
- Snapshots on intents or submissions record what price applied for later audits.
Operators maintain categories, rules, campaigns, and payment-method toggles in the operator console.
Review depth by submission family
SIGMA does not treat every submission type the same when it prices review depth:
- SKILL / manifest-led submissions lean mostly on manifest scope, declared capabilities, mismatch signals, and other lightweight trust-surface indicators.
- API-bearing submissions lean mostly on endpoint surface: how many endpoints are declared, how many are authenticated, and how many are mutating (
POST/PUT/PATCH/DELETE). - CODE-bearing submissions lean mostly on repository footprint and risky behavior: scoped file count, total source size, dependency count, language spread, and higher-risk execution behavior.
In mixed categories, the heavier surface dominates:
- Skill + API: endpoint surface matters more than manifest prose.
- Skill + Code: code footprint matters more than manifest prose.
- Skill + API + Code: code and API surface dominate; the skill/manifest acts as a lighter modifier unless it introduces mismatches or risky declarations.
This means a small SKILL.md with a large codebase should price like a code review, while a tiny code skill with many endpoints should price more like an API review.
Package-aware categories
Package-aware categories add a separate dependency certification layer on top of the main submission quote:
- Packages are charged per uncertified direct dependency.
- Each package is tiered from the downloaded package source footprint, not from the parent submission multiplier.
- Operator policy can define an optional total package-cost cap so dependency-heavy projects do not grow without bound.
Developers should therefore read the quote as two layers:
- Submission review depth for the main skill / API / code review.
- Package audit cost for dependency certification when a
...+Packagescategory is selected.
Manifest generation
AI manifest assist is rate-limited and can be locked after repeated abusive input; the UI points you to Support when that happens.
Validator SIGMA plugins (SBYTE)
Separate from developer deposits: creating SIGMA-type validator plugins debits SBYTE from the validator’s game wallet to the protocol treasury after balance checks and an on-chain transfer. Other plugin families may be free in the same pricing system.
Ledger and model costs
SIGMA tracks LLM usage for protocol accounting. Live model price sync and every billing detail vary by deployment—use operator LLM usage and finance views for reconciliation.