General Reasoning is designed for autonomous systems whose populations may grow far beyond human-seat licensing. We therefore separate access to the GR platform from the volume of governed work the platform processes.
A commercial deployment combines a platform entitlement with GR Usage Units (GRUs), a standardized unit of governed activity. Chandra Core provides the basic mechanics of Chandra inside the customer's environment. It preserves the local evidentiary record from which governed-work usage can be reconciled without transferring customer evidence. GR commercial products are proprietary and distributed under a commercial source license. Licensed customers receive the Lisp source required to operate, inspect, compile, debug, and administer the authorized environment; source delivery does not transfer ownership or commercialization rights.
AletheiaGraph, Chandra Enterprise, Chandra Passport, or DXMachine. Each tier includes the dependencies beneath it. GRIC is included throughout.
Qualified customer infrastructure or an AegisGenera appliance. The software dependency model is the same in either deployment.
Commercially meaningful governed activity consumes GR Usage Units. The usage ledger is reconciled against Chandra evidence rather than an opaque proprietary counter.
A customer enters the stack at any valid dependency boundary. The named tier is the highest capability licensed; required dependencies are included automatically.
| Commercial tier | Included |
|---|---|
| AletheiaGraph | AletheiaGraph + Chandra Core + GRIC |
| Chandra Enterprise | Chandra Enterprise + AletheiaGraph + Chandra Core + GRIC |
| Chandra Passport | Chandra Passport + Chandra Enterprise + AletheiaGraph + Chandra Core + GRIC |
| DXMachine | DXMachine + Chandra Passport + Chandra Enterprise + AletheiaGraph + Chandra Core + GRIC |
A Context Unit is not a token, credit, or tradable object. Once created, it is immutable evidence of a specific attributed event — effectively frozen in carbonite.
Commercial schedules define standardized units of governed work for settlement. Those settlement quantities are derived locally from customer evidence and do not require Context Units to leave the organization that created them.
| Context Unit (CU) | GR Usage Unit (GRU) | |
|---|---|---|
| Purpose | Immutable evidence | Commercial capacity |
| Fungible | No | Yes |
| Tradable | No | Potentially transferable under agreement |
| Created by | Chandra Core / governed runtime | General Reasoning entitlement system |
| Consumed by | Never; retained as evidence | Governed commercial activity |
A governed action does not have one universal economic value. A clinical recommendation, a securities settlement decision, an energy dispatch instruction, and a routine administrative update may all produce Context Units, but they represent different levels of consequence, governance burden, regulatory exposure, and enterprise value.
The commercial model therefore does not assume 1 CU = 1 GRU. Context Units remain the universal evidence primitive. Sector policy interprets the work described by those CUs and determines how much commercial capacity that work consumes.
A work class can incorporate clinical consequence, protected health information, authorization requirements, human review, provenance, and whether the result is advisory or directly consequential.
A work class can incorporate transaction value, fraud exposure, monetary consequence, dual approval, regulated recordkeeping, settlement finality, and external execution.
A work class can distinguish planning, forecast reasoning, asset control, dispatch instructions, grid-impacting state changes, and other consequential operational actions.
Lower-consequence workflow, file, records, and business-process activity can use a different schedule without forcing every enterprise action into the economics of highly regulated sectors.
The objective is not to create a separate currency for every industry. GRUs remain the common settlement unit. Sector-specific schedules determine the number of GRUs consumed by a classified unit of work.
| Evidence layer | Classification layer | Commercial layer |
|---|---|---|
| Context Units Universal, immutable evidence of attributed activity. | Sector work class The semantic and consequential class of work represented by that evidence. | GR Usage Units The fungible commercial capacity consumed under the applicable schedule. |
ILLUSTRATIVE CLASSIFICATION CU evidence: sector: healthcare work-class: clinical-recommendation consequence-class: high authority-class: physician-approval data-class: PHI external-effect: false Sector schedule: Healthcare / clinical recommendation Commercial result: GRU consumption determined by the contracted healthcare work schedule.
AletheiaGraph can participate in this classification by reasoning over the context surrounding the work: principal, role, workflow, data classification, target asset, transaction band, approval path, CRC surface, consequence class, and other authoritative facts. The resulting classification can remain inspectable in GRIC rather than becoming an opaque billing label.
Once governed work is classified consistently, an enterprise can reserve or hedge the capacity it actually expects to need. A financial institution might contract for a forecast volume of transaction-governance work, while a health system contracts for a different mix of clinical reasoning and consequential clinical actions. Both settle in GRUs, but each operates under the applicable sector work schedule.
Chandra Core provides the basic mechanics of Chandra inside the customer's environment. It creates and preserves the local evidentiary record used by higher-level GR products and by the customer's own operational and audit processes.
General Reasoning does not need customer Context Units, graph content, business records, event timing, principal identities, workflow details, or application payloads in order to reconcile commercial usage.
Commercial reconciliation is based on agreed aggregate work classifications and cryptographic commitments derived locally from the customer's own evidence. The evidence itself remains under customer control.
Context Units, graph content, business data, event timing, and operational detail remain inside the customer's environment. Market participants see contract state and settlement quantities, not customer runtime activity.
For the Chandra evidence model and protocol mechanics, see chandraprotocol.org.
Context Units belong to the organization whose governed environment created them. They remain inside that organization's evidentiary domain. General Reasoning does not require customer CUs, graph content, business records, application payloads, principal identities, exact event timing, or workflow detail for commercial settlement.
ILLUSTRATIVE SETTLEMENT STATEMENT Organization: Example Health System Period: 2028-Q1 Schedule: HEALTH-2028-v3 Clinical H1: 18,204,112 Clinical H2: 41,887,340 Clinical H3: 7,291,883 Clinical H4: 191,044 Evidence commitment: cryptographic commitment to the private evidence set Underlying Context Units: retained by the customer not transmitted to General Reasoning
Customers purchase GRUs directly from General Reasoning as included capacity, prepaid capacity, or additional consumption blocks.
Customers maintain the commercial entitlement for the GR software tier they operate, independent of individual human or agent counts.
Integrators, resellers, or counterparties may enter approved contracts referencing future governed-work capacity or pricing without taking custody of customer Context Units.
Where contracts are transferred or novated, the shared settlement system records the contract-state change while the customer's underlying evidentiary records remain private.
Autonomous workloads can scale unpredictably. Large organizations may want price certainty without prepaying every possible future unit of work. The commercial model is therefore designed to support capacity commitments, forward contracts, and, for qualifying enterprise agreements, options on future standardized governed work.
An organization commits today to a defined quantity of standardized governed work for a future period at agreed commercial terms. GR gains committed demand; the customer gains cost predictability.
A customer pays a premium for the right, but not the obligation, to settle a defined quantity of future governed work at a specified strike price during an agreed exercise window. The underlying work is measured locally when it occurs.
As the market matures, enterprise agreements may support additional structures such as collars, novations, or standardized futures. The initial design preference is to keep these instruments tied to measurable governed work rather than an independently speculative token.
The usage statement should be inspectable in GRIC and reconcilable against customer-local Chandra evidence. Customers should be able to understand what classes of governed work were counted, while GR receives only the agreed settlement quantities and cryptographic commitments needed to reconcile the contract.
ILLUSTRATIVE MONTHLY CAPACITY STATEMENT Starting GRU balance 50,000,000 Received 2,000,000 Transferred 5,000,000 Consumed 17,428,331 Remaining 29,571,669 Consumption classes Healthcare H2/H3 8,204,114 Financial F2/F3 5,110,080 General enterprise 4,114,137 Classification Work classes are derived from CU evidence under the applicable sector schedule. Evidence Usage is reconcilable to Chandra records through Chandra Core.
GR applications are implemented as live Lisp systems. Licensed customers therefore receive the proprietary source required to operate, inspect, compile, load, debug, administer, and maintain the GR software included in their authorized capability tier.
Source delivery is intentional. It does not constitute a sale or transfer of the software, its source code, or General Reasoning intellectual property. The customer receives contractual rights to possess and use that source only within the scope of the applicable agreement.
An initial, non-refundable source-entitlement fee compensates General Reasoning for granting access to the proprietary implementation included in the licensed tier. It is separate from governed-work usage and is not credited against usage unless the applicable agreement expressly says otherwise.
A minimum annual commercial commitment establishes meaningful production participation. The commitment is financial rather than operational and can be applied against governed-work settlement during the contract term.
Qualifying work is reconciled under the applicable sector and work schedule. Usage above the committed amount is settled according to the agreement.
COMMERCIAL STRUCTURE
Proprietary source entitlement
consideration for source access and internal operation
+
Annual minimum commercial commitment
creditable against governed-work settlement
+
Governed-work settlement
actual qualifying usage under the agreed scheduleSubject to the applicable agreement, source access may permit internal inspection, compilation, loading, debugging, operation, and authorized internal modification of GR software within the licensed environment. Support obligations may require an issue to reproduce against an unmodified qualified GR release.
Standard source entitlement does not permit publication, redistribution, sublicensing, sale of copies, disclosure to unauthorized third parties, incorporation into a competing commercial product, or operation as a third-party commercial service outside the licensed scope. Any OEM, embedded, redistribution, hosted-service, or re-commercialization rights require a separate written agreement.
General Reasoning commercial software is licensed for production use under a proprietary commercial source entitlement. Production access requires a minimum annual commercial commitment appropriate to the licensed capability, deployment scope, support level, and applicable governed-work schedule.
The minimum commitment is a financial commitment, not a workload quota. Amounts paid toward the commitment are applied against governed-work charges during the contract term according to the applicable commercial agreement. Usage above the committed amount is settled under the applicable work schedule.
A customer is never required to manufacture transactions, Context Units, autonomous actions, or artificial workload simply to retain a production entitlement. If actual governed-work usage is below the committed amount, the commercial obligation is still satisfied by the agreed minimum spend rather than by generating unnecessary activity.
Time-limited or otherwise restricted non-production access can be offered for evaluation, development, integration, and qualification.
Production entitlement requires the applicable minimum commercial commitment and permits operation within the licensed customer and deployment scope.
Larger commitments can support organization-wide deployment, negotiated sector schedules, support, AegisGenera fleets, and future-capacity agreements.
A production entitlement grants the contractual right to operate the specified GR software within the customer, deployment scope, term, and other conditions stated in the agreement. It does not transfer ownership of the software, source code, intellectual property, or commercialization rights.
Redistribution, sublicensing, resale, hosted third-party commercialization, and incorporation into another commercial software offering are not included in a standard production entitlement. Those rights, if available, require a separate written OEM, embedded, redistribution, or other commercial agreement with General Reasoning.
An AegisGenera deployment combines the physical appliance with the selected GR software entitlement and governed capacity. The hardware does not create a separate pricing universe.
| AegisGenera configuration | Commercial software included |
|---|---|
| AegisGenera / AletheiaGraph | AletheiaGraph + Chandra Core + GRIC |
| AegisGenera / Chandra Enterprise | Chandra Enterprise + AletheiaGraph + Chandra Core + GRIC |
| AegisGenera / Chandra Passport | Passport + Chandra Enterprise + AletheiaGraph + Chandra Core + GRIC |
| AegisGenera / DXMachine | DXMachine + Passport + Chandra Enterprise + AletheiaGraph + Chandra Core + GRIC |