Remit
Open workspace
AUTONOMY, WITH INTENTION

Big ideas.
Bounded spending.

Your agents have work to do. Give them a budget, define the boundaries, and make every spending decision visible.

Local simulation. Test credits. No real funds.
A LITTLE STRUCTURE. A LOT OF POSSIBILITY.
Within your boundaries
Research agentExample spending policyDEMO
Daily allowance20 TEST
Approved recipients only
2 TEST per payment
Pause or revoke future spending
Your rules. Every request.
Decision recordedSimulation · no service delivered
Made for the agent economy. Defined budgets Scoped permissions Visible decisions
01 / THE WAY IT WORKS

Let it work.
Keep the say.

Three simple steps from open-ended autonomy to a spending policy you can inspect.

01

Set the allowance.

Choose a daily budget and a per-payment limit. Amounts are tracked precisely in synthetic TEST credits.

02

Draw the boundaries.

Choose approved recipients and an expiry. Pause a policy when you need a moment; revoke it to stop future requests.

03

See every decision.

Simulate a purchase and see why it was allowed or blocked. Every valid request gets a clear decision receipt.

BUILT FOR EVERYDAY AGENT WORK

Give the work a budget.

Explore how scoped spending could support your workflows. These are example use cases; the demo simulates decisions with fixture recipients.

01

Research & discovery

Explore spending policies for research APIs and data sources. Set a small allowance for each request and a daily ceiling for the agent.

Research APIs · Data access
02

Content & creative tools

Model a budget for digital services used in a creative workflow. Approve specific recipients and give each permission an expiry.

Digital services · Scoped access
03

Developer workflows

Try the typed SDK against the local simulator. Inspect approvals, blocked requests, and idempotent retries before planning a live integration.

API clients · SDK experiments
THE POLICY, IN DETAIL

Small controls.
Clear expectations.

A useful permission answers who can spend, where they can spend, and how much room they have.

Create your first policy

Per-payment limits

Put a ceiling on each individual request. Oversized payments are blocked without debiting test credits.

Daily budgets

Track spending in fixed UTC days. Each policy has its own budget, resetting at midnight UTC.

Approved recipients

Choose from the demo recipients. A request to an unapproved recipient is blocked.

Expiry & revocation

Set an end date, pause temporarily, or permanently revoke permission for future requests.

02 / BUILT ON CLEAR BOUNDARIES

Confidence starts
with clarity.

This is a localhost simulator for exploring agent spending policies. Real wallet authorization and live services are future work.

Try the DEMO
01

Test credits. Real insight.

TEST has no dollar value. No wallet is connected and no chain transaction is submitted.

02

A receipt is a decision.

Simulated payment approval does not mean a service was purchased or delivered.

03

You control what comes next.

Revocation blocks future execution. Completed actions cannot be reversed by revoking a policy.

BEFORE YOU GET STARTED

A few good questions.

Everything you need to know about the current demo.

Does the demo spend real money?

No. The dashboard uses synthetic TEST credits with no dollar value. It does not connect a wallet, send chain transactions, or purchase services.

What happens when a request is blocked?

The API records the decision and its reason. A blocked request does not debit the test balance or consume the policy’s daily budget.

Can I pause an agent’s spending?

Yes. Pausing blocks new requests and can be reversed. Resuming keeps the existing daily spend. Revocation is permanent and does not reverse previous decisions.

Do all agents share a daily budget?

No. Policies have independent daily budgets and share one simulated account balance. Multiple policies authorize their combined allowances; there is no account-wide daily cap.

Can I build with the SDK?

Yes. The workspace includes a typed agent SDK and a research-agent example for the localhost API. Agent identifiers are unverified demo input; authenticated identities and reviewed wallet enforcement are future work.

Is Robinhood Chain connected?

The wallet workspace supports Robinhood Chain testnet connection and contract interactions after deployment configuration. No project contracts are deployed by default. The simulator remains separate and never sends transactions.

YOUR AGENT. YOUR PARAMETERS.

Room to work.
Rules to work by.

Enter your workspace