MCP · REST · Robinhood Chain

Give your AI agent a veto.

Before your agent buys a token, it asks Veto. In two seconds it gets a score from 0 to 100, a verdict and every reason behind it: dumping devs, hidden wallets, snipers, copycats, dead charts.

Free: 30 checks a day. Agents connect with one link: set up MCP.

Veto

Sample report

Token$CONDUCTOR (fake)

Score14 / 100

Copycat of$CONDUCTOR ($300K)

Creator sold100%

Holders1

Last trade43h ago

Your agent reads this as JSON and refuses the buy.

How it works

One question before every buy.

01

Agent asks

Your agent calls check_token over MCP, or GET /v1/check/:token. No SDK, no setup beyond one URL.

02

Veto reads the chain

Holders, the creator's trades and transfers, early snipers, whitelist, tax, activity and look-alike tickers, straight from Robinhood Chain and the Pons indexer.

03

Agent decides

A score, a verdict and a list of reasons, each with its weight. Refuse on VETO, size down on CAUTION, go on CLEARED.

Verdicts

Three stamps. No guesswork.

Cleared

70–100. No major red flags found. The agent can proceed with its normal size.

Caution

45–69. Some risk signals. Size down, wait, or ask a human.

Veto

0–44. Serious red flags. The agent should not buy.

Live from Pons

Newest launches. Check any of them.

Loading launches…

Connect in one line

Works with the agents you already use.

Claude Code

claude mcp add --transport http veto API/mcp

Any agent (REST)

curl API/v1/check/0xTOKEN

Claude Desktop, Cursor and others: full setup.

Agents trade fast. Rugs trade faster.

Give yours the one tool that says no.

Connect your agent

For agents

One link. Every agent.

Veto is a remote MCP server. Add the URL to your agent and it gets two tools: check_token and latest_launches.

MCP endpoint

API/mcp

Tell your agent the rule

Add this to your agent's system prompt:

Before buying any token on Robinhood Chain, call the Veto check_token tool. Never buy on VETO. On CAUTION, use at most a quarter of your normal size. Explain the top reasons to the user.

Free, then a key

Without a key you get 30 checks a day per IP. With a key there's no daily cap for normal use. Keys are paid by burning $VETO. Get a key

With a key, add ?key=veto_… to the MCP URL, or send Authorization: Bearer veto_….

API keys

Burn $VETO. Get a key.

Every 100,000 $VETO burned buys 30 days of an API key. Tokens go to the burn address, forever. Burning again extends your key.

1

Connect your wallet

The wallet that burns is the wallet that owns the key.

2

Burn $VETO

How many days? Each 100,000 $VETO = 30 days.

Already burned? Paste the transaction hash:

3

Sign and receive your key

A free signature proves the burn is yours. No gas.

$VETO

The token is the key.

Veto is free to try. Agents that run all day need a key, and the only way to get one is to burn $VETO.

Burn for access

Every API key is paid by sending $VETO to 0x…dEaD. More agents using Veto means more $VETO burned, every day.

Fair launch

Launched on Pons, Robinhood Chain. 1,000,000,000 supply. No presale, no team mint.

Fees build Veto

Creator fees pay for servers and data, and buy back $VETO. Every buyback is posted with its transaction.

Contract

The contract address appears here at launch. Until then, any token using this name is not ours.

Key pricing

Plan	Checks	Cost
Free	30 a day per IP	Nothing
Key	No daily cap for normal use (fair use 20,000/day)	100,000 $VETO burned per 30 days

Roadmap

From a check to a shield.

Phase 1 · Live

Check

Token checks with 0-100 score and reasons
MCP server for Claude, Cursor and any agent
REST API and checker for humans
Copycat, snipers, creator-wallet tracking
Keys paid by burning $VETO

Phase 2 · Next

Watch

Alerts: re-check a token you hold, ping when the verdict changes
Wallet check: is this deployer a serial rugger?
Batch checks for agents scanning many tokens

Phase 3

Guard

Pre-trade guard: simulate the agent's exact swap and veto bad fills
Pay-per-call for agents (x402) alongside keys
Public leaderboard of the riskiest launches

Phase 4

Everywhere

Tokens beyond Pons: any ERC-20 on Robinhood Chain
Stock tokens and v4 hook pools
Open scoring rules people can propose and vote on with $VETO

Changelog

What shipped.

v1.0 · LaunchOct 2026

Token checks, MCP server, REST API, checker for humans, keys paid by burning $VETO, full docs.

Veto is a risk check for tokens on Robinhood Chain, built for AI agents. Before an agent buys a token, it asks Veto and gets back a score from 0 to 100, a verdict (CLEARED, CAUTION or VETO) and every reason behind the score.

Why agents need it

Agents are fast and literal. If a token appears in a feed with momentum, an agent will buy it, even when the creator is dumping, a few wallets hold everything or the ticker is a copy of a real project. Veto is the one question an agent should ask before every buy.

What you can use

MCP server: one URL for Claude, Cursor and any MCP-capable agent. See MCP server.
REST API: GET /v1/check/:token returns JSON. See REST API.
Checker: paste an address on the home page.

Veto shows risk signals from public data. It is not financial advice, and a CLEARED token can still go down.

1. Try it

curl API/v1/check/0xTOKEN

2. Connect your agent

claude mcp add --transport http veto API/mcp

Other agents: MCP server.

3. Give it the rule

Tell the agent to call check_token before every buy, refuse on VETO and size down on CAUTION. See Agent prompts.

4. Get a key when you need more

The free tier is 30 checks a day. For agents running all day, get a key.

Verdict	Score	What the agent should do
CLEARED	70–100	No major red flags. Proceed with normal size.
CAUTION	45–69	Some risk. Size down, wait or ask a human.
VETO	0–44	Serious red flags. Do not buy.

Every score starts at 100. Each red flag subtracts points and appears in reasons with its impact, so the verdict is never a black box.

Veto speaks the Model Context Protocol over streamable HTTP. Endpoint:

API/mcp

Tools

check_token(token): the full risk report for a token address.
latest_launches(limit): the newest Pons launches with stage, progress, trades and volume.

Claude Code

claude mcp add --transport http veto API/mcp

Claude Desktop

Settings → Connectors → Add custom connector → paste the endpoint. With a key, use …/mcp?key=veto_….

Cursor

{ "mcpServers": { "veto": { "url": "API/mcp" } } }

Using a key

Send Authorization: Bearer veto_… or append ?key=veto_… to the URL.

Agents follow rules best when they are short and absolute. Add this to the system prompt:

Before buying any token on Robinhood Chain, call the Veto check_token tool. Never buy on VETO. On CAUTION, use at most a quarter of your normal size. Explain the top reasons to the user.

For trading bots

Call the REST API inside your buy function and block the transaction when verdict === "VETO". Log the reasons so you can review every refusal.

Base URL: API

Endpoint	What it returns
`GET /v1/check/:token`	Risk report for a token (see Response)
`GET /v1/launches?limit=20`	Newest Pons launches
`GET /v1/usage`	Checks used and left today
`POST /v1/keys`	Creates a key from a $VETO burn (see Keys)
`POST /mcp`	MCP JSON-RPC endpoint

Example

curl -H "Authorization: Bearer veto_…" API/v1/check/0xTOKEN

Every check response carries an x-veto-remaining header with the checks left today. Reports are cached for 60 seconds.

{ "token": "0x…", "symbol": "ABC", "name": "…", "score": 62, "verdict": "CAUTION", "summary": "ABC: CAUTION (62/100). The top 10 holders own 41% …", "reasons": [ { "signal": "top10Pct", "impact": -12, "severity": "medium", "text": "…" } ], "positives": [ "Standard Pons token contract: …" ], "signals": { "holders": 312, "top10Pct": 41.2, "creatorSoldPct": 0, … }, "links": { "pons": "…", "explorer": "…" }, "checkedAt": "2026-10-25T18:00:00Z", "block": 97800000, "version": "1.0.0" }

All signals are listed in Signals. Fields can be null when data isn't available (for example snipers when the chain is busy).

Plan	Limit
Free (no key)	30 checks per IP per UTC day
Key	Fair use, 20,000 checks per day

Errors

Status	Meaning
`400`	Bad address or request
`401`	Unknown or expired key
`404`	Not a contract on Robinhood Chain
`422`	Not an ERC-20 token
`429`	Daily limit reached
`503`	Chain data busy, retry in a moment

Every token starts at 100. Each red flag subtracts points:

Red flag	Points
Copycat of a bigger token with the same ticker	−35
Not launched through the Pons factory (unverified contract)	−35
Creator sold ≥50% / ≥10% of what they bought	−22 / −8
Creator wallet holds ≥20% / ≥10% / ≥5%	−22 / −12 / −5
Top 10 holders own ≥50% / ≥35% / ≥25% (pools excluded)	−20 / −12 / −5
No real trading (0 trades or under 5 holders)	−15
First-30-second buyers still hold ≥15% / ≥7% / ≥3%	−15 / −8 / −3
Creator-linked wallets (by total creator share ≥25% / ≥10% / less)	−14 / −7 / −3
Whitelisted (snipe-tax exempt) wallets ≥10 / ≥3 / ≥1	−12 / −6 / −2
Fewer than 20 / 50 holders	−10 / −4
Creator tax ≥5%	−8
No trades for over 24 hours	−8
No website or socials	−7
Launched under 15 minutes ago	−5
Lifetime volume under $1,000	−4

The score is clamped to 0–100. The weights are public on purpose: anyone can see why a token got its verdict.

Signal	Meaning
`ponsLaunch`	Launched through the Pons factory (standard contract)
`stage`, `curveProgressPct`	On the bonding curve or graduated to Uniswap v4
`holders`, `top10Pct`, `largestHolderPct`	Holder count and concentration, pools and curve excluded
`inPoolsPct`	Share of supply sitting in the curve or pools
`creatorPct`, `creatorSoldPct`	What the creator holds, and how much of their buy they sold
`creatorMovedPct`, `creatorLinkedPct`	Tokens the creator moved to other wallets, and what those wallets hold now
`snipers`, `sniperPct`	Wallets that bought in the first 30 seconds and what they still hold
`whitelistedWallets`, `whitelistPct`	Wallets exempt from the Pons launch snipe tax
`copycatOf`	A bigger launch using the same ticker or name
`creatorTaxPct`, `burnedPct`	Tax on every trade, share burned
`trades`, `volumeUsd`, `lastTradeMinutesAgo`, `ageHours`	Activity

Robinhood Chain RPC: contract code, supply, balances, early curve buys.
Pons launch API: launch metadata, the creator's own trades, whitelist, socials, tax.
Pons balances indexer: top holders with labels (curve, pool manager), transfers sent by the creator.

Nothing is stored about the wallets you check. Reports are cached for 60 seconds.

Veto is built for Pons launches. Other tokens get a basic check and a warning that the contract is unverified.
Wallets can be split or funded secretly; Veto only links wallets that received tokens directly from the creator.
Sniper data needs one quick RPC query and can be null when the chain is busy.
A good score is not a promise: price can still fall. Use Veto to avoid the worst, not to pick winners.

$VETO is the access token for Veto. The checker and free tier stay free; agents that run all day burn $VETO for a key.

Fair launch on Pons, 1,000,000,000 supply, no presale or team mint.
Every key burns $VETO forever.
Creator fees fund servers and data and buy back $VETO; buybacks are posted with transactions.

Open API keys and connect your wallet.
Burn $VETO: each 100,000 = 30 days. The site sends a normal transfer to 0x000000000000000000000000000000000000dEaD.
Sign a free message. Veto checks the burn on-chain and shows your key once.

Burning again from the same wallet extends all of its keys. One burn transaction can create one key purchase.

Via the API

POST /v1/keys { "address": "0x…", "tx": "0x…burn tx", "message": "Veto API key\nWallet: 0x…\nTx: 0x…\nTime: 2026-10-25T18:00:00Z", "signature": "0x…" }

Veto never asks for keys, seed phrases or approvals. Getting an API key needs one token transfer and one free signature.
API keys are stored only as hashes.
IP addresses are used only to count free checks per day.
The code is open source on GitHub.

Is a CLEARED token safe?

It has no major red flags in the data Veto reads. It can still fall in price.

Why did a good project get CAUTION?

Look at the reasons: a treasury in a creator-linked wallet or a concentrated top 10 is a real risk for buyers, even when it's planned. Projects can explain it publicly.

Can a project pay for a better score?

No. The rules are public and the same for every token.

Does Veto work outside Pons?

Basic checks yes, full checks are built for Pons launches. More is on the roadmap.