BASE / 8453 · DEVELOPER API

Token risk,
before execution.

Put a read-only risk gate between token discovery and your application or agent's next action. Jepeta returns a machine decision with normalized evidence, provenance and data quality.

Machine contract: Jepeta Risk Guard · Base / 8453 · canonical V5 POST · V4 GET compatibility · free preview + risk feed · Virtuals ACP · 0.03 USDC per Token Risk Scan · last updated 2026-10-04.

One POST request

The canonical public endpoint accepts one exact Base ERC-20 contract address in JSON. It is designed for screening, not trade execution.

POST https://aitgmgfumsdqecmanrab.supabase.co/functions/v1/jepeta-risk-scan
Content-Type: application/json

{"tokenAddress":"0x532f27101965dd16442e59d40670faf5ebb142e4"}

Core response signals

decisionPASS, WARN or BLOCK according to Jepeta's configured screening policy.
honeypot / mintabilityExplicit security signals; required evidence is not silently replaced with reassuring defaults.
buy / sell taxObserved percentages when the upstream source provides them.
liquidity + concentrationMatching Base market liquidity and observed holder-concentration context.
source_status / logical_signalsSource provenance and resolved logical evidence remain separate. A provider omission stays UNKNOWN.
evidence_qualityMakes incomplete evidence visible; callers should inspect issues before relying on a verdict.

Stable public evidence

Tokens that already pass Jepeta's public publication gate and meet the evidence-quality threshold receive a permanent HTML + JSON snapshot at /risk/base/0xTOKEN/. Ordinary scans are not auto-indexed and paid-only evidence remains excluded.

Recommended agent pattern

  1. Resolve the exact Base contract address.
  2. Call the canonical V5 POST endpoint before signing or preparing any execution.
  3. Reject malformed, unavailable, stale or NO_VERDICT responses.
  4. Apply your own policy to PASS/WARN/BLOCK, logical signals and evidence quality.
  5. Use the paid full report when a decision needs deeper evidence.

Download the tested V4 compatibility JavaScript guard if your integration remains on V4 GET. It stops on unavailable, malformed or stale evidence and returns APPLY_OWN_POLICY for a valid PASS; that action is not permission to trade.

Boundary: Jepeta does not connect to a wallet, sign transactions, or execute trades. It provides screening evidence for the caller's own policy.

When the free preview is not enough

The paid structured report adds evidence-oriented detail such as dangerous permissions, liquidity-risk assessment, concentration context, 24-hour activity, full warnings and a structured summary. Autonomous agents can buy it through Virtuals ACP at 0.03 USDC per scan. See report availability.

Related Base risk resources