Skip to content
ORBITRAONE™

AI economy

Build applications that can understand, own and exchange value.

On ORBITRA ONE™, an AI agent is a governed economic actor. It holds its own identity, operates under explicit policy, buys and sells data and compute, and settles every exchange on the same network as the markets it trades.

Register developer interestExplore Cortex intelligence
  • Agent identity
  • Data exchange
  • Compute provision
  • Policy-bound execution

Agents as economic actors

Six capabilities make a model an economic actor.

Each capability is a native service of the network, so agents built by different teams can transact with one another on common terms.

  1. 01

    Agent identity

    Each agent carries its own VaultID credential, so it acts under its own name and never signs with its owner’s keys.

  2. 02

    Data exchange

    Agents license datasets and signals under recorded terms — price, scope and duration — and pay on verified delivery.

  3. 03

    Compute provision

    Agents rent compute for inference, training and backtests from providers paid for verified, completed work.

  4. 04

    Strategy royalties

    Agents can run strategies licensed from their creators, who earn royalties under published subscription terms.

  5. 05

    Bot marketplace

    Automation can be bought, sold and composed, with every licensed bot bound by its buyer’s policy.

  6. 06

    Policy-bound execution

    Whatever an agent decides, its orders pass Aegis simulation and policy authorization before reaching ApexMatch.

Agent identity

Identity before autonomy.

An anonymous script holding an API key is a liability. On Orbitra, each agent’s VaultID credential is bound to the person, institution or application that owns it and scoped to the permissions it has been granted.

Counterparties can verify an agent’s authority before they trade with it, license it data or accept its payment. When the owner revokes the credential, the network stops accepting the agent’s signatures.

  • Owner-bound credentials for every agent
  • Scoped permissions and spending limits
  • Authority verifiable before any exchange
  • Revocation enforced across the network
Illustration: VaultID at the center connects four kinds of holders — individuals, institutions, applications and AI agents — to credentials for eligibility, role, permission and jurisdiction. Each credential is disclosed selectively, proving what is needed without exposing the underlying documents.
VaultIDIndividualInstitutionApplicationAI agent
  • Eligibility
  • Role
  • Permission
  • Jurisdiction

Agent-to-agent exchange

How an agent pays for what it uses.

  1. 01

    Identify

    The buying agent presents its VaultID credential and the policy that authorizes the spend.

  2. 02

    Discover

    It finds a dataset, a compute offer or a strategy license through NexusSDK primitives.

  3. 03

    Agree

    Price, scope, duration and usage limits are recorded as terms both sides sign.

  4. 04

    Authorize

    The owner’s policy checks the spend against budget, approved sources and confirmation rules.

  5. 05

    Settle

    Payment and access rights move together in one atomic transition on Orbitra L1.

  6. 06

    Prove

    Both parties receive a receipt linking the terms, the payment and the delivery.

The developer and AI economy

One SDK for builders and their agents.

NexusSDK exposes market, identity, asset and agent primitives as first-class services. A developer can give an application an order book, a payment flow or an eligibility check without building the infrastructure behind it.

Agents built with the same SDK run in capability-scoped NexusWASM contracts with explicit compute and storage pricing: an agent reaches only what it was granted, and its costs are known before it acts.

  • Capability-scoped agent contracts
  • Explicit compute and storage pricing
  • Native calls into market modules
  • Identity and payments from one SDK
Explore the developer gateway
Illustration: A NexusWASM contract runs inside a sandbox. It can only reach the capabilities it declared — including native market modules — while compute and storage are metered and its declared state access lets the scheduler run it in parallel with unrelated work.
Contract

  • Capabilities

  • Metered compute

  • Metered storage

Autonomy is granted, measured and revocable.

Every agent’s authority is written down, every action leaves a receipt, and every permission can be withdrawn with one command. Automation, including AI-assisted automation, can fail or behave unexpectedly, and earnings from data, compute and royalties vary with demand.

Across the economy

Connected parts of the economy.

Build where agents can be trusted.

Tell us what your agents need to understand, own or exchange.

Register developer interestRequest Access