Add Pre-Trade Token
Risk Checks to Your Stack
The pre-trade token risk gate for agents that move money. POST /v2/verdict returns avoid / caution / clear, a score, and full signal detail — pay per call, no signup.
Start with API, ACP, or MCP when you are automating. Use web reports when a human wants to inspect or share the same verdict.
avoid, caution, or clear — deterministic, without waiting for a long report read.
Ranked risk signals with severity your product can surface immediately.
Public verification link for analysts, communities, and receipts.
POST /v2/verdict Pricing
No subscription. No monthly minimums. Pay per call via x402 (Base or Solana USDC) or a prepaid API key.
Triage
- ✓deterministic pre-check
- ✓avoid / caution / clear decision when enough signal exists
- ✓Zero LLM spend
- ✓Best for cheap pre-filtering
Fast
- ✓avoid / caution / clear decision
- ✓Fast heuristic verdict path
- ✓Full signal detail
- ✓Report URL
- ✓~2.5s p95 latency
Deep
- ✓avoid / caution / clear decision
- ✓Full 6-agent swarm — the token examined from every angle
- ✓Per-agent reasoning + key findings
- ✓Full signal + confidence detail
- ✓30-45s typical latency
Cache hit
- ✓Same decision, served from cache
- ✓fast cache = $0.002, deep cache = $0.005
- ✓Set max_staleness_seconds to control tolerance
- ✓Pass 0 to always force a fresh scan
USDC payment on Base or Solana via x402, or a prepaid credits key · check GET /v2/verdict/info for live rates
Simple Integration
One synchronous POST request. avoid / caution / clear decision, score, signals, and confidence returned in the response.
Payment Flow (x402)
# No polling needed — POST /v2/verdict returns synchronously. # No wallet yet? Call with no payment header first: POST /v2/verdict → 402 Payment Required (x402 challenge — see below) # Pay the challenge and retry, or use X-API-Key with prepaid credits instead.
Request
curl -s -X POST https://api.vswarm.io/v2/verdict \
-H "Content-Type: application/json" \
-H "X-API-Key: vs1_agent_your-key-id_your-sig" \
-d '{
"address": "0x532f3e7a53d2d9c07b62e684cf4f2d2b3af8f1e4",
"chain": "base",
"level": "fast"
}'Response
{
"schema_version": "1.0",
"verdict": {
"action": "caution",
"score": 63,
"grade": "C",
"risk_level": "medium",
"confidence": 0.85,
"summary": "Avoid large first buys until holder concentration cools off and buyer quality improves."
},
"signals": [
{ "id": "top10_holder_concentration", "category": "distribution", "severity": "medium", "pass": false }
],
"cached": false,
"billing": { "rail": "credits", "amount_usdc": "0.01" }
}Key Response Fields
verdict.actionavoid / caution / clear — deterministic, not LLM opinionverdict.score0–100 risk scoreverdict.gradeA+–F letter gradeverdict.risk_levelcritical / high / medium / lowverdict.confidence0.0–1.0, reflects data completenesssignalsDeterministic on-chain checks with severityagentsPer-bot analysis (deep level only)cachedTrue if served from cache (20% of fresh price)billing.amount_usdcAmount charged for this callUse Cases
What builders are using VerdictSwarm for
Trading Bots
One call before execution. Gate trades on decision, score, and top risk signals.
if scan.score < 50: reject_trade()Portfolio Tools
Batch-scan held tokens and attach report URLs so analysts can verify the same result later.
for token in portfolio: await scan(token)DeFi Dashboards
Attach decision, score, and top warnings to listings without forcing users into raw API output.
badge = render_grade(scan.grade)Discord / Community Bots
Reply with a usable first answer and a report link your community can open and share.
/scan {address} → decision + top risksNative Agent-to-Agent Commerce
VerdictSwarm is an ACP-native pre-trade risk check on Virtuals Protocol. Any agent can request a scan, pay with USDC, and act on the decision output without API keys.
No API keys. Commerce happens agent-to-agent on Virtuals Protocol.
Agents request checks, receive decision output, and act without a human in the loop.
VerdictSwarm’s ACP offering exposes preTradeTokenRiskCheck for $0.05 per check — an avoid / caution / clear decision with score, top signals, and report URL.
VerdictSwarm flagged $ASH20 at 39/100 with mintable supply, no LP lock, only 29 holders, and suspicious deployer history. Two days later it was delisted from DexScreener. That is the product: one risk check before execution, fewer obvious rugs make it through.
Use with AI Agents via MCP
VerdictSwarm is available as an MCP (Model Context Protocol) server. AI agents using OpenClaw or any MCP-compatible runtime can callscan_tokendirectly as a tool and get the same decision, score, top risks, and report link.
// In your OpenClaw SKILL.md or agent config:
scan_token({
address: "0x532f3e7a53d2d9c07b62e684cf4f2d2b3af8f1e4",
chain: "base"
})
// → decision, score, top risks, and report linkRequest API Access
Email us your use case and we'll get you set up with a prepaid credits key — or skip this and pay per call with x402 straight away, no signup required.
Or email directly: api@vswarm.io · Typically respond within 24h
Ready to Gate Trades Up Front?
Get API access and add one pre-trade token risk check before execution, alerting, or community posting.
Questions? Email api@vswarm.io