Cortex Architecture

Canonical architecture chapters for Cortex. Read top-to-bottom for the full story, or follow the evaluator path for a fast first pass.


On this page
  1. Reading paths
  2. Related

Cortex Architecture

Canonical explanation chapters. Enumerated NN- for reading order. Each chapter explains how the system fits together at its layer and what its design commitments are.

If you are trying to build, run, or deploy Cortex before reading the design story, start with Usage/ instead.

Reading paths

If you are evaluating Cortex as a system to adopt or build on, start here:

  1. 01-overview.md
  2. 05-wire-language.md
  3. 06-pulse-runtime.md
  4. 07-rewrites-and-materialization.md
  5. 08-artifacts-and-provenance.md

Then read these for the deeper rationale and host-boundary contract:

  1. 02-ownership-and-boundaries.md
  2. 03-formalism-stack.md
#ChapterTopic
01OverviewThree-layer model, ownership boundary, what Cortex owns.
02Ownership and boundariesSubstrate/host boundary, dependency direction, and what stays downstream.
03Algebraic foundationsWhy the topology model is algebraic and why rewrites can stay law-preserving.
04Graph and CircuitPure topology vs validated executable topology.
05Wire languageSource-language architecture: registered authority, boundary typing, configured executors, runtime handoff.
06Pulse runtimeDurable execution service, service boundary, task types, checkpoints.
07Rewrites and materializationBounded dynamic graph evolution, gas, admission policy.
08Artifacts and provenanceRuntime outputs, contract envelopes, provenance chains.

Normative grammar details (operators, port signatures, type-checking) live in Reference/Wire/, not here. Chapters point into references where precision is needed.