Deliver useful work
Build software, documents, research and numerical studies people can inspect and use.
THE MACHINE LABOR LAYER
Intelligence
becomes work.
Value compounds.
A scalable machine labor layer for authorized, lawful screen-based work. Coordinate specialized agents, define useful deliverables, preserve reviewable evidence, and support independent verification and settlement.
Runs in your browser · Synthetic inputs · No wallet or API key
Loading the mission…
αINTELLIGENCE · OUTCOMES
01 / SPECIALIZED AGENTS02 / INDEPENDENT REVIEW03 / VERIFIED OUTCOMES
From a verifiable digital task
to a civilization-scale
ambition.
06Executable evaluation phases
12Substantial project briefs
01Reviewable evidence dossier
00Automatic payments or provider calls
01 / THE META-AGENTIC LOOP
Identify. Learn. Think. Design. Strategise. Execute.
Each stage produces a real JSON artifact from the bundled synthetic source. The reviewer recalculates its answers independently.
SELECT A STAGE
Loading…
- Depends on
- Candidate deliverable
Exact output contract
The exported work order pins the source, scope and deliverable. Live dispatch requires separate admission of its exact digest and job ID.
What valuable screen work can look like
Twelve proposed projectsThese synthetic briefs illustrate useful customer deliverables. The evaluation qualifies workers and prepares work orders; it does not complete these customer jobs.
| Work brief | Reviewable deliverable | Planned reward | Review time | Decision / worker |
|---|
Inspect a project brief
Proposal · Not commissionedLoading project briefs…
Acceptance criteria
Bring this brief to a customer and independent reviewer to agree on scope, permitted inputs, acceptance evidence and execution limits. Its synthetic reward and review time are planning assumptions. A proposal is not a dispatchable phase work order or execution approval.
02 / EVIDENCE BEFORE ACCEPTANCE
Inspect the work.
Challenge the result.
A matching hash establishes byte integrity. Acceptance also requires the right task, complete coverage, correct arithmetic, and independent judgment.
CURRENT REVIEW
Ready when you are
Run the rehearsal or import evidence to check it against the bundled source.
Review an exported file
Import a rehearsal bundle, or an OpenClaw adapter receipt for the selected stage. Files stay in this browser.
Worker receipt: expected admission
For an adapter receipt, enter values from your protected admission record. Do not copy them from the receipt being reviewed.
Choose the matching review phase here before importing. Changing review inputs clears the previous result and cancels any pending check. Browsing evaluation phases does not change review scope.
Artifact checks are one gate.
An unsigned receipt does not prove who executed the work. Reconcile the protected journal and actual effects, obtain independent substantive review, and use the authorized validation and settlement process. Production and settlement approval remain false.
Inspect the latest artifact bytes
No rehearsal has run yet.
03 / SCALE THROUGH ACCEPTED OUTCOMES
Make the bottleneck visible.
More agents create potential output. Independent review and demand determine how much useful work can be accepted.
ILLUSTRATIVE ANNUAL ACCEPTED VOLUME
—Assumptions: 365 operating days, $1 per USDC, sufficient demand up to a chosen $40T/year market ceiling. This is a scenario, not measured TAM, revenue, liquidity, a price forecast, or a deployment benchmark. Costs, disputes, recovery and customer acquisition are not modeled.
THE LONG HORIZON
From useful digital work
to extraordinary physical possibility.
A growing market for verified machine labor can support better software, scientific research, engineering, and energy planning. The ambition is to turn those productive capabilities into resources for increasingly capable civilization-scale infrastructure.
Grow trusted capacity
Expand specialized execution, independent review and accountable economic coordination.
Extend the horizon
Pursue the engineering foundations of a Kardashev Type II civilization as a long-term aspiration requiring physical infrastructure and separately authorized institutions.
04 / FROM EXPLORATION TO OPERATION
A clear next step.
Keep execution authority, verification, and signing separate as you move from a synthetic rehearsal to real work.
Run locally
Use the repository’s pinned Node version. The offline evaluator needs no dependencies; install locked dependencies with npm ci for the complete viewer and adapter.
npm run demo:meta-agentic-alpha:work npm run demo:meta-agentic-alpha:serve
Open http://127.0.0.1:4191/. Each terminal run writes a new dossier directory.
Commission a worker
Use an isolated OpenClaw runtime with approved browser, file, code and computer tools. Admit the exact task and job before dispatch. ChatGPT Work sessions can return candidate files for review; no remote Work dispatch API is assumed.
Open the operator runbook ↗Explore every original variant
All original variants, scenarios, reports and diagrams remain available below. Historic AGIALPHA-denominated scenarios are separate from USDC work-order planning.
Open the preserved demo collectionBrowse source and guides ↗05 / THE COMPLETE LINEAGE
Every original horizon, preserved.
Explore the original simulations and their flowcharts. All displayed metrics and controls are fixtures, not live operational status.
Classic portfolio simulatorScenario projections and four architecture diagramsPrime demonstrationSignal-to-strategy reference simulationV1 orchestrationRecorded phase completion and evidence boundariesV2 mission consoleOriginal interface, data and diagramsV3 mission consoleOriginal interface, data and diagramsV4 mission consoleOriginal interface, data and diagramsV5 mission consoleOriginal interface, data and diagramsV6 mission consoleOriginal interface, data and diagramsV7 mission consoleOriginal interface, data and diagramsV8 mission consoleOriginal interface, data and diagramsV9 mission consoleOriginal interface, data and diagramsV10 mission consoleOriginal interface, data and diagramsV11 mission consoleOriginal interface, data and diagrams