The Runcible Product Stack

Resources / Architecture

The Runcible Product Stack

An examination language, a falsification engine, an enterprise operating platform, and the record that connects them.

Revised 10 November 2026

In this article
  1. RDL: make the claim explicit
  2. Runcible AI: examine the necessary conditions
  3. Oversing: run the institution
  4. The Decidability Record: preserve the examination
  5. The research and improvement foundation

Runcible’s product architecture connects an examination language, a falsification runtime, and an enterprise operating platform. The layers share a record of actors, operations, evidence, obligations, and consequences, while retaining different responsibilities.

RDL: make the claim explicit

The Reality Description Language is the representation layer for semantic computation. It identifies the entities, relations, operations, conditions, evidence requirements, interests, and limits needed to examine a claim.

Ordinary prose can hide a missing actor or an unsupported causal step inside a persuasive paragraph. Operational representation makes those absences visible. A contract assertion, a resource forecast, and a policy proposal retain their different evidentiary burdens while becoming examinable through a common grammar.

RDL is a continuing formalization and compiler-development program. Operational prose and implemented protocols supply the existing working foundation; progressively stronger representations improve precision, repeatability, and compilation performance.

Runcible AI: examine the necessary conditions

The runtime subjects candidate claims to constructive falsification. It reconstructs what must be possible, derives relevant examinations, applies protocols, uses evidence and tools where required, and records surviving, defeated, and unresolved conditions.

The four burdens remain distinct: possibility, ethics and reciprocity, truth and testifiability, and warrantability and liability. The engine also exposes errors and concealed costs and proposes repairs or settlement options. The runtime’s output is an examined record, not a grant of corporate authority.

Oversing: run the institution

Oversing is the Semantic ERP: an enterprise operating platform for continuous adaptation. Its scope connects executive objectives and organizational structure to people, customers, commitments, delivery, finances, and responsibility.

  • Strategy and organization: Goals, reporting relationships, locations, lines of business, and executive reporting.
  • People and capability: Recruiting, skills, staffing, payroll, and continuing employee development.
  • Commercial commitments: CRM, professional services automation, pricing, contracts, and budgets.
  • Operational delivery: Programs, projects, queues, tasks, dependencies, and scheduling.
  • Financial consequence: Currency integration, accounting, allocation, balances, and real-time reporting.

Its permissions and procedures govern decisions and authorized execution. Runcible and the generating model can recommend; Oversing issues the permitted task to a separate executor. The action and its consequences return to the institutional record.

Current implementation: The runtime, protocols, retained examinations, and Oversing application exist. The inspected saved-conversation pilot blocks AI business mutations. Bounded delegated execution and complete cross-system action-and-outcome linkage are release work. The separation described above defines the architecture those releases must implement and demonstrate.

The Decidability Record: preserve the examination

The common artifact retains the question, construction, evidence, test results, repairs, limits, and open dependencies. Oversing connects it to authority, work, and outcomes. Different audiences can receive different presentations of that record without silently changing the underlying findings.

A stored record is useful only to the extent that its sources, tests, and execution can be inspected. A valid schema or cryptographic digest establishes neither the truth of an input nor the sufficiency of the examination by itself.

The research and improvement foundation

The Natural Law research and book program supplies the underlying work on measurement, testimony, cooperation, and evolutionary computation. RDL and protocols translate that work into explicit representations and examinations. Examined cases contribute to a governed Truth Corpus and to training and evaluation material.

The research program is organized across six volumes:

  1. The Introduction — the Natural Law of Decidability.
  2. The Crisis of the Age.
  3. Language as a System of Measurement.
  4. The Logic of Evolutionary Computation.
  5. The Science of Human Behavior.
  6. The Law — a formal logic of decidability applied to law.

The volumes provide explanation, reference material, and specifications for the development of protocols and processes. This index of the research program does not imply that every volume or working manuscript is publicly released.

The immediate development path is incremental: improve RDL, compiler behavior, protocols, and control over existing models; reduce repeated work and inference cost; evaluate each change against retained cases. More specialized model training and changes to learned representations are further engineering paths, each with its own evidence requirements.

Read the Architectural White Paper for the coupled-system design and the Oversing overview for the operating platform’s full scope.