A Veto-guarded agent holding a Veto card

For your agents

Give your agents money. Keep control.

Your agents can pay for what they need online - APIs, tools, data, purchases - from a wallet you control. Veto authorizes every spend against your policy before it happens, so they can't overspend, can't wander off-limits, and every payment is signed.

Install the CLI Read the docs

Works with Claude Code, Codex, Cursor and any agent.

One command

Install Veto into your agent.

One line installs Veto and wires its policy tools into your agent. From then on, every payment your agent tries to make is checked against your policy first - no extra code to write.

Works with Claude Code, Codex, and Cursor. Bring your own agent, or start from veto-agents - a ready-made agent with Veto already inside.

or auto-detect

# Claude Code

$ npx -y @veto-protocol/cli@latest claude

How it works

Autonomy, inside a boundary you set.

You hold the keys and set the rules. Veto sits in front of every spend and enforces them - in real time, and on-chain.

Connect a wallet

Bring any wallet. You hold the keys; Veto only holds a co-signer.

Set your policy

Caps, allowlists, time windows, categories - in plain terms.

Fund it

Run veto wallet fund for a QR, then send USDC on Base.

The agent tries to spend

Before any payment it calls veto_authorize. Veto checks the intent against your policy.

Allow, deny, or escalate

Within policy it pays. Out of bounds it's blocked - or escalated to you. Every decision is signed.

Money with no brakes.

The agent holds the private key outright
Nothing checks a payment before it goes out
One bad prompt or bug can drain the wallet
No caps, no allowlist, no way to say no
You find out what it spent after the fact

Money with a policy.

You hold the keys - Veto only holds a co-signer
Every spend is authorized against policy first
Caps and allowlists enforced on-chain, not on trust
Out-of-bounds spends blocked in real time
A signed, verifiable receipt for every decision

Your policy

You decide the rules. Veto enforces them.

Set the boundary in plain terms - no smart-contract knowledge required. Change it any time; the new version takes effect on the next spend.

Spend caps

Per-transaction, daily, and monthly caps.

Allowlists

Approve the services your agent can pay.

Time windows

Restrict spending to business hours.

Rate limits

Cap how many payments happen per hour.

Signed receipts

Every decision signed and reason-coded.

Give your agents money - safely.

Set the policy once. Veto enforces it on every spend, on-chain - so you get autonomy without handing over a blank cheque.

Install the CLI Read the docs →

Non-custodial - you hold the keys On-chain hard-stops Signed receipts