Core
Foundational semantics for shapes, observations, values, expressions, annotations, and the shared vocabulary other Cohesive blocks use.
The Cohesive Building Blocks
Cohesive is organized into semantic blocks that define one system graph: shapes, relations, transitions, processes, integrations, presentation, identity, storage, simulation, infrastructure, and the runtime bindings that interpret them.
Blocks
Relations
Define typed relationships, projections, and queries once, then evaluate them across supplied objects, databases, indexes, and composed sources.
Entities
Define business entities, invariants, and typed Transitions that return explicit decisions, state changes, domain events, and requested work.
Processes
Coordinate entity changes, requests, events, waits, parallel work, and longer-running business workflows using asynchronous C#.
Integrations
Declare bounded ingestion responsibilities and compile acquire, publish, ledger, and settlement boundaries into canonical Processes.
Presentation
Backend-owned UI semantics for views, actions, forms, flows, data bindings, and generated frontend contracts.
Configuration
Typed configuration semantics for profiles, dependency selection, environment overlays, validation, and runtime composition.
Identity
Identity, authorization, roles, principals, permissions, and access semantics modeled independently from any specific auth provider or transport.
API
API semantics for operations, routes, request shapes, response variants, scope policies, transport metadata, and generated clients.
Storage
Entity repositories, canonical query sources, durable Process checkpoints, and Relation-derived materialization with explicit capabilities.
Simulation
Deterministic synthetic data, property cases, replayable worlds, retained artifacts, and provider-neutral provisioning for tests and tools.
AI
AI and numerical-computation semantics for inference, training, datasets, vectors, text processing, and model artifacts.
CLI
Command-line application semantics backed by typed configuration, host-backed dependency injection, executable tools, and local harnesses.
CodeGen
Code generation tooling for projecting the system graph into concrete artifacts while keeping identifiers, types, routes, selectors, and contracts aligned.
Infra
Portable requirements, capability-qualified realizations, lifecycle ownership, physical readiness, and adapter boundaries.
Building-block map
One graph, carried from model to application
Model blocks state the system. Integrations lower bounded ingestion into Processes, Storage binds requirements to physical media, Infra proves a coherent realization, and project blocks expose the graph.
Core, Configuration, and AI are cross-cutting semantics. The Model category contains Relations, Entities, Processes, and Integrations. Relations may be independent or connect entity shapes and transitions. Entities define invariants and state transitions. Processes coordinate queries, transitions, effects, and multi-entity workflows. Integrations lower bounded ingestion into canonical Processes. A highlighted path carries model requirements into Storage, then into storage and interactions and an Infra realization. Identity sits in the Realize category because placement and enforcement happen there, while dashed connections show that authority spans the model and its projections. CLI, API, and UI project the graph to people and tools. Simulation provides assurance across the model, realization, and projected behavior.
Cross-cutting
Semantics available throughout the graph
Model
State what the system means
bind semantic requirements
Realize
Place and execute the model
Storage
Maps model requirements to persistence and interaction capabilities
Storage + interactions · repositories · indexes · streams · queues
Infra
Proves target capability coverage and records physical realization
project the same graph
Project
Expose the graph to people and tools
Assurance
Replay decisions, verify bindings, and inspect projected behavior