lookx402 methodology · glossary · faq · api

About lookx402

Independent public observatory for x402 protocol activity on Base. This page covers who we are, what we deliberately do not do, how we handle data, and how to reach us. Updated 2026-05-11.

What lookx402 is

lookx402 is a real-time public observatory for AI-agent payments made via the x402 protocol on Base mainnet. We index every observable x402 transaction, decode it correctly (extracting the EIP-3009 authorizer rather than the facilitator EOA), aggregate by agent and merchant, classify behavior into nine archetypes, and expose the result as web pages and a public JSON API.

The site went live on 2026-04-27 and runs continuously since then, ingesting Base mainnet RPC every 5 minutes. Full pipeline at methodology.

Who we are

lookx402 is an independent project. It is not a company, not a foundation, and not affiliated with any token or treasury. There is no team page because the project is run by a small group of contributors who prefer to be reached through the project's public X account rather than individually. Editorial responsibility rests with that account.

Domain expertise: blockchain indexer engineering (Base / Ethereum L2 / EVM event log decoding), behavioral analysis of on-chain wallets, public-data infrastructure on Cloudflare and Supabase, and protocol-specific reverse engineering of EIP-3009 / EIP-712 / Permit2 settlement flows.

Editorial policy

We treat lookx402 as a public-interest reference rather than a commercial product. Three principles:

1. On-chain truth over narrative convenience.

If the data says the agentic economy is concentrated in one wallet, we say so — even when it makes a less flattering headline than a hypothetical "thousands of bustling agents" story.

2. Pseudonymity is a property to respect, not a puzzle to solve.

We label wallets only when their owner has self-disclosed the address on a surface they control (their website, their docs, their GitHub README, an ENS reverse record). We do not chain-cluster wallets to natural persons.

3. Methodology is the product.

Our differentiation versus generic explorers is not branding — it is correct decoding of EIP-3009 authorization events. The methodology page documents the full algorithm. If you can replicate it with the same RPC strategy and rule table, you should arrive at the same numbers.

What we deliberately do not do

No funds flow through us. lookx402 holds no hot wallet, no treasury, no custodial balance, no token. We are a passive read from public RPC.

No facilitator role. We do not broadcast EIP-3009 authorizations, do not sign anything on anyone's behalf, do not relay payments. Inspecting how facilitators behave is part of our methodology — performing facilitation is not.

No paid labelling. Merchants and agents cannot pay to be added to, removed from, or reordered in our leaderboards.

No paywall, no rate limit, no key. The JSON API is fully public and unauthenticated.

No clustering to natural persons. We do not deanonymize wallets. We do not attribute wallets to companies absent verifiable self-disclosure.

No fingerprinting of visitors. We do not run third-party analytics that fingerprint, do not run ad pixels, do not store visitor PII beyond standard, anonymized 30-day server logs. We log crawler User-Agents (Googlebot, GPTBot, ClaudeBot, etc.) for indexer-health reasons only.

Data handling & RGPD posture

lookx402 indexes only public on-chain data. The data we surface — wallet addresses, transaction hashes, USDC amounts, block timestamps — is by definition public the moment it lands on Base mainnet.

We treat wallet addresses as entity identifiers, not person identifiers. We apply the self-disclosure-only labelling rule above.

If you control a wallet and want a self-declared label removed from lookx402, ping @lookx402 on X. We remove without question.

If you believe a label on lookx402 is inaccurate, contact us through the same channel. We treat corrections as critical and patch promptly.

Standard server logs (request path, timestamp, anonymized IP block, User-Agent) are kept for 30 days for indexer-health and abuse-detection purposes, then deleted. No personal accounts, no cross-site tracking.

Corrections policy

We treat methodology-level errors as critical. If you find a decoding error, a misclassification, a stale label, a broken link, or a factual mistake on any page:

  1. Ping @lookx402 on X with the URL and a short description.
  2. We acknowledge within 48 hours.
  3. We fix or respond with a reason within 7 days.
  4. Methodology changes are logged in the methodology changelog with a dated entry.

Conflicts of interest, advertising, and sponsorship

PracticeOur position
Display advertisingNone.
Sponsored contentNone. We have published nothing sponsored and will not in the future.
Affiliate linksNone.
Paid placement in leaderboardsNone. Order reflects on-chain volume only.
Token / treasury ownership of x402 ecosystem assetsNone held by the project. Contributors as individuals may hold USDC as is normal in this space.
Grant / fundingSelf-funded. If that changes we will publish the source here.
Personal stake in any agent or merchant we indexIf a contributor controls an indexed wallet, they recuse from any editorial decision about that wallet. As of writing, no such overlap exists.

Independence statement

lookx402 has no affiliation with:

The absence of affiliation is what lets us call methodology errors when we see them, including in tools published by the entities above.

How to cite lookx402

Citation rules and BibTeX block live at methodology · how to cite. Short version: cite as lookx402 (https://lookx402.com) with the access date and a link to the specific page or API endpoint that produced your number.

Contact

Best channel: @lookx402 on X. We monitor mentions and DMs and respond in English or French.

For methodology questions or corrections: same channel — please include the URL and a precise description.

For press inquiries: same channel — we are happy to help with quotes, charts, or methodology explainers.

This page is licensed CC0 — anyone can reuse the editorial-policy language verbatim if it helps establish similar standards on another public-data observatory.