ALPHA AGI MARK / CAPITAL TO CAPABILITY

Give ambition
a working engine.

Turn a mission into funded intent, bounded computer work and evidence someone else can verify. Build useful capacity. Earn the next step.

Plan a mission ↓Inspect the contract demo ↗

Interactive planning simulation · No sign-in or wallet · Files stay in this tab until downloaded

Two connected ideas. Separate execution.

The original Alpha Mark demo explores a native-currency or configured ERC20 bonding-curve market, launch gates and refunds. This lab prepares a USDC-denominated work proposal. Neither browser calculations nor a market launch establish that useful work has been completed.

FROM A THESIS TO A USEFUL RESULT

Capital needs a purpose.
Work needs proof.

01 / DEFINE

A mission with boundaries.

Choose a buyer, approved public or licensed inputs, explicit permissions and a deliverable with measurable acceptance criteria.

02 / EXECUTE

Specialists, coordinated.

A commissioned OpenClaw worker or operator-led ChatGPT Work session can use its available browser, desktop, file and code tools within the admitted scope.

03 / VERIFY

An independent judgment.

Preserve artifacts and source hashes, reproduce the checks and record each acceptance decision. Reserve reviewer time before scaling the worker pool.

04 / SETTLE

Value the buyer accepts.

Only an authorized deployment can release funds under its actual rules. Buyer use, review and settlement evidence determine what really happened.

THE CAPITAL-TO-WORK LAB

Scale the work.
Keep the review.

Change the assumptions and see the limiting resource.
All outputs are drafts and modelled capacity.

Browser-only planning. No source is fetched, worker connected, transaction signed or money moved. Do not enter private information or credentials. Preparation checkboxes record your intentions; they grant no authority.

1. Define useful work

Try an operator-guide job for this public repository. Replaces the work definition and clears preparation acknowledgments; keeps your capacity assumptions.

Work category Objective Scope and measurable acceptance detail Approved source URLs

One public HTTPS URL per line, up to 20. No credentials, query strings or fragments. Availability and rights need separate operator verification.

Input class

Execution route

Worker profile name

The operator must commission and admit the actual profile. This name does not connect a worker.

2. Model one day of capacity

Daily reward budget (USDC)

Reward per job (USDC)

Unfunded reward proposals only. Each amount must be between 0.000001 and 1,000,000 USDC, with up to six decimals; excludes provider charges, review fees, rework and disputes. Verify deployment currency separately.

Workers

Candidate jobs per worker / day

Independent reviewers

Minutes per reviewer / day

Run allowance per job (minutes)

Review allowance per job (minutes)

Run allowance is exported for operator planning; it does not calculate worker throughput or enforce runtime limits. Jobs per worker is a separate assumption.

3. Prepare the operator handoff

Confirm these preparation commitments before exporting the proposed task. At least one complete job must fit your capacity assumptions.

I will verify the buyer’s authority, permitted actions and exclusions. I will verify source rights, approved origins and exclusion of private data. I will assign a reviewer independent of the worker and agree acceptance criteria.

CAPACITY IS A CONSTRAINT

What can fit?

Enable JavaScript to calculate a local draft.

0candidate job slots / day

Static upper bound under your assumptions. Not a success, delivery or revenue forecast.

Worker capacity: 0
Review capacity: 0
Reward budget capacity: 0

How the model works

Worker slots = workers × assumed jobs per worker. Review slots = reviewers × floor(minutes per reviewer ÷ minutes per job). Each complete job has one independent reviewer; partial minutes from different reviewers are not pooled. Reward slots = floor(daily reward budget ÷ reward per job). The smallest count bounds candidate slots. Failed attempts, review overhead and buyer demand can reduce useful output further.

Inspect the complete draft JSON

What the buyer receives

A reproducible analysis, numerical results and environment record.

What the reviewer checks

Pin source data and the computational environment; specify units and numerical tolerances.
Have an independent reviewer rerun the calculation and explain any discrepancy.

The existing adapter accepts bounded UTF-8 artifacts. Office files, screenshots, binaries and large applications need a separately commissioned artifact store. The downloaded task is compatible with the existing work planner’s schema; it still requires operator admission.

Inspect actual delivery evidence →

CHOOSE THE NEXT CONCRETE STEP

Explore. Reproduce. Commission.

THE MARKET DEMO

Inspect the launch gates.

Reproduce capital formation, policy checks, launch and refund scenarios on a disposable local chain. Review reports as evidence of those local scenarios.

Guided tour and original diagrams →

npm run test:alpha-agi-mark
npm run demo:alpha-agi-mark:full

Use the pinned Node/npm toolchain and npm ci first. Read the runbook before configuring an external network.

THE WORKER ROUTE

Connect authorized execution.

OpenClaw executes through the repository’s admitted Responses adapter and a protected worker profile. ChatGPT Work is an operator-led route with its own tools and permissions; no remote Work dispatch API is exposed here.

Worker integration guide →Open the complete work planner →

THE EVIDENCE ROUTE

Make readiness observable.

Verify target-network configuration, trusted release, actual provider calls, independent review, buyer use and authorized settlement. This planning site supplies no production certification.

Current readiness index →Complete delivery workflow →

A LONGER HORIZON, BUILT FROM USEFUL WORK

From one verified result
to extraordinary capacity.

Coordinate software, science, energy and infrastructure research into larger programs. The ambition is cumulative: better tools enable better experiments, stronger evidence and more capable teams. Physical infrastructure outcomes require engineering and validation beyond a screen.

$40T / year

User-supplied scenario assumption for broad screen-based labor, not a verified market estimate, revenue forecast or claim that all human tasks can be automated. Available tools, permissions, task reliability and independent review determine achievable work.

Explore the civilization-scale research model →

Give ambitiona working engine.

Capital needs a purpose.Work needs proof.