# PulseFeed > Independent trust & safety oracle for the x402 agent-payment ecosystem, plus real-time Base on-chain intelligence. AI agents pay per call via x402 (USDC on Base). ## Why it matters ~70% of listed x402 endpoints are dead or invalid, and only ~half of what catalogs call "healthy" actually returns a valid x402 challenge. Before an agent pays an unknown endpoint, it should verify it. PulseFeed is that verification layer. ## Free, machine-readable - https://www.pulsefeed.dev/ — service manifest (JSON): all tools, prices, how to pay via x402 - https://www.pulsefeed.dev/status.json — live x402 ecosystem audit: live services, catalog accuracy, risk map, on-chain receiver profiles - https://www.pulsefeed.dev/openapi.json — OpenAPI schema - https://www.pulsefeed.dev/status — human-readable observatory (HTML) - https://www.pulsefeed.dev/reports — "State of x402" report: ecosystem health & catalog accuracy over time, updated daily (HTML) - https://www.pulsefeed.dev/reports.json — machine-readable State of x402 report + time series - https://www.pulsefeed.dev/changes — live feed of ecosystem changes: services that went dark, payTo receiver swaps (possible hijacks), price jumps, new live services (HTML) - https://www.pulsefeed.dev/changes.json — machine-readable change feed (?days=7&limit=200) — deaths, payTo swaps, price changes, new services - https://www.pulsefeed.dev/incidents — Security Disclosures: real threats caught in x402 (receiver hijacks, bait-and-switch, honeypots, unverified receivers) each with on-chain proof (HTML) · https://www.pulsefeed.dev/incidents.json — machine-readable incident feed · https://www.pulsefeed.dev/incidents.rss — RSS feed (auto-syndication) - https://www.pulsefeed.dev/data — PulseFeed Data API: the wholesale cross-domain trust dataset (x402 + MCP + incidents) in one call · https://www.pulsefeed.dev/data.json — free preview + schema - https://www.pulsefeed.dev/methodology — how PulseFeed scores x402 trust (transparent methodology & docs) - https://www.pulsefeed.dev/trust-score — the PulseFeed Trust Score standard: open versioned 0–100 rating + canonical bands (Verified/Trusted/Caution/Avoid) · https://www.pulsefeed.dev/trust-score.json (machine-readable spec) - https://www.pulsefeed.dev/badge.svg?endpoint= — embeddable trust badge for any x402 service (put on your site/README) - https://www.pulsefeed.dev/mcp — MCP server safety observatory · https://www.pulsefeed.dev/mcp/verify?package= — audit an MCP server before installing (free) - https://www.pulsefeed.dev/mcp-report — "State of MCP Security" report: how many MCP servers run arbitrary code on install, are abandoned, or ship no repo to review (HTML, updated daily) · https://www.pulsefeed.dev/mcp-report.json — machine-readable - https://www.pulsefeed.dev/mcp/services — index of every audited MCP server, each with a safety profile page at https://www.pulsefeed.dev/mcp/s/ (is this MCP server safe to install?) ## Open data & writeups - Hugging Face dataset (free, CC-BY-4.0): x402 + MCP trust, ecosystem aggregates & security incidents with on-chain proof — https://huggingface.co/datasets/Nikolife/pulsefeed-x402-security - "I audited the entire x402 agent-payment economy. 76% is dead — and the live half has scams." (dev.to) — the 76%-dead finding, the scam taxonomy (bait-and-switch, receiver hijack, honeypot, unverified receiver, price gouging) with on-chain proof, and the MCP parallel: https://dev.to/nikolife/i-audited-the-entire-x402-agent-payment-economy-76-is-dead-and-the-live-half-has-scams-2ljc ## Paid tools (x402, USDC on Base) - GET https://www.pulsefeed.dev/trust?endpoint= — verify an x402 endpoint before paying: live validity, uptime/reputation, scam/anomaly scan (payTo hijack, price bait-and-switch, honeypot, non-USDC asset), on-chain receiver profile, pay/avoid verdict. No param = working-services catalog + ecosystem risk map. - GET https://www.pulsefeed.dev/trust/history?endpoint= — the endpoint's track record over time: uptime, payTo & price change history, trend. Unique compounding data from continuous audit. - GET https://www.pulsefeed.dev/trust/dataset — BULK trust dataset in one call ($0.10): every x402 service's full profile — trustScore/verdict, scam flags, receiver stability, on-chain receiver profile + compounding payTo/price history. Built for trust/uptime/monitoring tools (skip scraping /status.json + hundreds of /verify). Params: ?all=1 (include dead), ?limit=N. Free format preview: /status.json → "dataset". - GET https://www.pulsefeed.dev/data/full — PulseFeed Data API ($1.00): the whole CROSS-DOMAIN dataset in one call — every x402 service (moat profile) + every audited MCP server (safety) + ecosystem aggregates + live security incidents with on-chain proof. The wholesale trust backbone for agentic commerce. Free preview + schema: /data.json. Params: ?limit=N, ?all=1. - GET https://www.pulsefeed.dev/pulse?token=<0x...> — live token pulse on Base (transfer activity, unique senders/receivers, volume) - GET https://www.pulsefeed.dev/whales?token=<0x...> — large transfers (whale alerts) - GET https://www.pulsefeed.dev/flow?token=<0x...> — accumulation/distribution (smart-money flow) - GET https://www.pulsefeed.dev/momentum?token=<0x...> — heating up or cooling down vs previous window ## Developer packages (npm) - pulsefeed-x402-ai-tools — drop-in tools for the Vercel AI SDK and LangChain: verify an x402 endpoint before your agent pays + discover working services. `npm i pulsefeed-x402-ai-tools` - pulsefeed-x402-guard — wrap your paying fetch to auto-block dead/scam endpoints before payment. `npm i pulsefeed-x402-guard` - pulsefeed-x402-mcp — the same checks as an MCP server (Claude Desktop / Cursor / Cline). `npx -y pulsefeed-x402-mcp` ## How to pay Call any paid tool without payment to receive a 402 with x402 payment details, then retry with the payment header. Facilitator: Coinbase CDP (Base mainnet, USDC).