Web page to clean text, title, links and metadata

Token security verdict (GET) — honeypot check, rug pull risk, scam signals, token safety grade. Query param: address=0x... Returns grade A-F + one-line reason + receipts. Re-call per token before each swap; verdicts change as owner/supply state changes.

Full token due-diligence report in one call — honeypot check, rug pull risk, buy/sell tax, mint and ownership privileges, blacklist/pause powers, holder concentration, LP lock status, DEX liquidity depth (Uniswap V2/V3 and Aerodrome), bytecode privilege scan, all fused into one graded verdict with per-signal receipts. Query param: address=0x... Use before a significant trade; cheaper than assembling five sources yourself.

Analyze French or English marketing and interface copy. Returns a clarity score, repetition data, sentence metrics, CTA detection and concrete fixes.

Counterparty check in one call. Send a company or person name, a domain, an EVM wallet address, or any mix; returns one report with a status per check: OFAC sanctions (name and wallet), SEC EDGAR and GLEIF company records (name), domain registration, DNS and TLS, and email authentication (domain); each check that returns data cites sources and retrieval times. Use before paying or onboarding a party. A screening aid, not a compliance determination. Failed calls are not charged.

Deepent Research — refusal-free LLM inference (per-request)

Token safety scanner for Base (JSON body): 0-100 safety score, risk band, and red flags for any Base token contract.

Use when an agent is building something Ethereum will hash and must get the same digest the chain gets: Keccak-256 of any string or hex payload, with the byte count. It is NOT SHA3-256, and that confusion is the usual reason a selector or a CREATE2 address comes out wrong. Do not use for file digests or API signatures outside Ethereum (/v1/util/sha256) or for a function selector, which is already built for you (/v1/util/selector).

Retrieve a public HTTPS URL from the US, Germany, Japan, Brazil, or India. Use HTTP mode for lightweight retrieval, or browser mode to render JavaScript pages. For pages with bot protection, set max_tier to L2 for managed unblocking at $0.05 USDC; L0/L1 costs $0.02 USDC. Returns the fetched body and an Ed25519-signed receipt for provenance.

Pre-trade safety check: oracle-immune-system verdict for one trade, with verifiable EIP-712 attestation support. BLOCK verdicts are complete checks and are charged.

Three-node GO/NO-GO threat intelligence report for an x402 Payment Required payload and vendor metadata on Base mainnet (eip155:8453). $0.05 USDC per POST via the CDP facilitator. Unpaid requests return HTTP 402 with PAYMENT-REQUIRED; a JSON body without a payload object returns 400; payment settlement failures return 402.

Real-time Base mainnet gas price signal (low/normal/high vs. recent average) — a recommendation on whether now is a cheap time to transact. Experimental/informational only — not financial advice.

Issue a signed read-only EVM integration-readiness report for a public address

Kronos BTC price forecast (latest, from the live HUD). Experimental/informational only — not financial advice.

Current Base mainnet gas price (wei/gwei), EIP-1559 base_fee_gwei/priority_fee_gwei, and gas_price_usd. Use when an agent must estimate gas cost in USD before sending a Base transaction.

Resize an image at a URL to a width, a height or both, keeping the aspect ratio inside the box by default (cover, contain, inside, outside or fill), optionally converting it. Returns a signed URL of the result, valid for an hour.

Audit real gas spend across up to 20 Base mainnet transaction hashes you provide — total/average gas used, cost in ETH, per-tx breakdown. POST body required. Experimental/informational only — not financial advice.

Simulate a Base mainnet transaction before submitting it — real eth_estimateGas + eth_call against the given {to, data, value}, returns whether it would succeed, estimated gas, and estimated cost. POST body required. Experimental/informational only — not financial advice.

Before buying on a French online store, check what an AI shopping agent found there: score out of 100, the six steps (access, cookies, search, product page, cart, checkout), whether checkout works without an account, shipping and returns facts with proof, and UCP checkout support. Dated tests with screenshots, public method.

Human approval for agents. Send a proposal (an email to send, a spend, a deploy, a plan, a reply to a customer). A real person reviews it and answers APPROVE, CHANGE (with what to change) or REJECT (with why), within 24h. A human reviewer's judgment, not a legal or financial authorization.