Documentation

Environment requirements, deployment steps, and a full walkthrough of a first task.

• 01
  OverviewWhat you need in place, the two ways to bring the platform up, and how to choose between them. Read this page first.
  3 min read
• 02
  DeploymentComplete steps for both deployment paths — local runtime and local Kubernetes — including prerequisites, private configuration and post-deployment verification.
  4 min read
• 03
  Your first taskA complete walkthrough from submitting a requirement to receiving a change request: what preflight does, what happens inside the sandbox, how artefacts are handed off, and how to confirm delivery succeeded.
  6 min read

The domain model behind tasks, flows, sandboxes, artefacts, knowledge and memory — and the reason for each constraint.

• 01
  Tasks and specificationsHow one R&D execution is defined, what is immutable while it runs, and why that immutability is the foundation of the product's reproducibility.
  4 min read
• 02
  Flows and work nodesHow R&D process is described as a versionable asset, how drafts differ from published versions, and why a tag alias is not a basis for reproducibility.
  4 min read
• 03
  Sandboxes and the execution modelWhat one-sandbox-per-task means, how the control plane and the sandbox divide responsibility, where state authority sits, and what this model costs.
  4 min read
• 04
  Preflight and freezingEverything that happens between task creation and sandbox creation: re-authorisation, deterministic compilation, capacity and risk assessment, digest computation, and the fail-closed boundary.
  4 min read
• 05
  Artefacts and unit gatesHow nodes hand off to one another, how artefact identity is composed, how unit gates decide, and why a downstream node cannot modify an upstream artefact.
  4 min read
• 06
  Knowledge and memoryTwo independent authority boundaries: how project knowledge reaches execution through human governance, and how personal memory is layered and frozen when a task is created.
  4 min read

How to create a task, observe execution, handle failure, and reuse capability.

• 01
  Create a taskRequirement sources, repository selection, flow mode and model freezing end to end — including the autofill behaviour and the order to triage a failed preflight in.
  5 min read
• 02
  Observe executionHow to read a task: the two-column structure, the disclosure order in node detail, the code and artefact evidence surface, and what a live connection actually means.
  4 min read
• 03
  Handle failures and recoveryTelling execution-level retry, semantic rework and manual rerun apart — what bounds each one, how far it reaches, and when to rebuild a task instead of recovering it.
  6 min read
• 04
  Flows and the capability marketplaceAuthoring, publishing and reusing flows and skills, how asset ownership is promoted, and why the runtime is never allowed to pull floating content.
  4 min read

Roles and permissions, deployment-level configuration, and how verification is graded.

• 01
  Roles and permissionsWhat each of the four perspectives can do, where those rights come from, why switching perspective never widens them, and the distinction between management target and current workspace.
  3 min read
• 02
  Platform configurationDeployment-level settings for the model gateway, Git services, sandbox policy, enterprise integrations and observability — and which of them belong to the platform rather than to a Workspace.
  4 min read
• 03
  Verification and acceptance boundariesThe four verification levels, and why none of them substitutes for another. This is the language any external capability claim has to use.
  3 min read

Canonical domain vocabulary and frequently asked questions.

• 01
  Domain glossaryThe canonical vocabulary for the platform's domain objects. Each entry gives a definition and the words that should not be used in its place, for aligning discussion, review and documentation.
  5 min read
• 02
  Frequently asked questionsCommon questions about product positioning, deployment shape, security boundaries and maturity, grouped by the kind of question asked.
  6 min read