FlywheelApp · full
Cinema + labs + WebGL. The labeled map, Hub canvas, field sketch, and every live engine. Use when this is the app that owns the GPU on the machine.Every feature, no graphics. Same engines, same API, same honesty grammar. No WebGL, no custom cursor, no Three.js. For when other apps already hold the GPU, or after learning loops when you only need the work.Terminal + one API out. Every live verb as a command and as HTTP. TUCE agents load skill.md and call /api/v1. No cinema overhead. Occupancy stays nominated | live-bound | unavailable.

Work · Train

Relearning reports here. Training does not start itself.

Training Lab holds recipes, preflight, start and cancel. The Observatory is read-only. Adaptation Lab dry-runs import. StrataX can report a completed job — that is a training report, not a hardware permission.

Training Lab

Start is a verb you press. Loss is not a score.

Loss and reward are optimization signals. They are not a product quality score and not a certification.

How it connects

Follow the occupancy.

documented

StrataX loops. Training Lab keeps the report.

A solve is a SAT result. It is not a hardware bind and not a training start.

  1. 01Objectivemodel · budget · constraints
  2. 02SAT resultfeasible or not
  3. 03Report to Trainjob, not a permission
  4. 04Observatoryread-only manifest

SAT ⟂ HAL. STDP is never a permission. If the remote harness is offline, local SAT remains the fallback. Loss is an optimization signal, not a quality score.

Three labs

Train, watch, adapt.

Lab

Training Lab

Recipes, preflight, start and cancel. Relearning reports land here.

Watch

Training Observatory

Read-only manifest: composition, provenance, holdout, leakage, ablation, resources, checkpoints.

Prep

Adaptation Lab

Dry-run import and certificate prep. Training stays off until you start it in the lab.

From StrataX

A completed loop is a report.

  1. 01

    StrataX runs a relearning loop with an objective, a model id, a loop budget, and constraints.

  2. 02

    SAT result: loops completed, feasible or not, engine, fingerprint.

  3. 03

    Report to Training Lab. SAT ⟂ HAL — a solve is not a hardware bind.

← ModelsOpen-source catalog. Fit before load. Download-only is not running.All integration flowsEvery Hub menu as a flow: scan, fit, bind, hot-swap, StrataX, observe, upgrade.Tools →Packs and plugins. Scan lists fingerprints; Accept is still required.