← Atlas ▌402atlas · report
The free onboarding report

x402, paid-tested: what's worth using.

8 services we bought with real USDC — each with the exact call and the real response. Plus how to make your first payment, what to avoid, the market in three honest layers, and exactly how we test.

updated 2026-06-04·// verify, don't trust·every claim has an on-chain receipt
01 — Make your first call

From zero to your first paid call

No account, no API key, no subscription. You need a wallet that can pay USDC on Base, and a client that speaks x402.

Wallet + USDC on Base

Coinbase Wallet or MetaMask on the Base network. Buy 5–10 USDC. Most services have a facilitator pay gas, so you can run on 0 ETH.

An x402 client

For a human at a terminal — the official CLI:

$ npx awal@latest

For an agent in code — the @x402/fetch library wraps fetch to pay automatically.

How a call works

request → 402 Payment Required → wallet pays USDC → retry → 200 + data. The client does pay-and-retry for you; it settles in seconds.

A real gotcha we hit (so you don't): the x402 spec moves fast — services now send the newer eip155:8453 network format, and older client libraries silently reject it and fail to pay. Use a current client (@x402/* v2.11+ or latest awal), not the old x402-fetch package. Keep your x402 wallet separate from your main funds — most services don't refund, so try a cheap endpoint first.
02 — What's worth using

The whitelist

8 services we paid for and got a real response from, grouped by what you'd use them for — not ranked by volume (in a wash-heavy market, high volume is the wrong signal). Every entry shows the exact endpoint and the actual body it returned.

Actions your AI literally cannot take

StablePhone ★ Merit Systems · the one we use most$0.54 call · $0.05 lookup
Your AI places a real phone call — dials a US number, speaks your task in natural voice, hangs up.
Endpoint
POST stablephone.dev/api/call · body: { "phone_number": "+1…", "task": "…" }
No model can place a phone call. Building telephony means numbers, carriers, compliance — not a weekend. $0.54 and it just dials.
Async: returns a call handle, then the AI dials seconds later. US numbers only.
paid-test record — how we bought it + on-chain receipt
Tested2026-05 · Merit Systems BoughtPOST stablephone.dev/api/call · body {phone_number, task} · $0.54 Receipt0xd7c5…c5d6c ↗ · Base · USDC

Market & price data — your AI can't know the live number

CoinGecko Pro ★ real brand · official x402$0.01
Live token prices & market data.
Endpoint
GET pro-api.coingecko.com/api/v3/x402/simple/price?ids=bitcoin,ethereum&vs_currencies=usd
What came back · $0.01 · snapshot 2026-06-04
{"bitcoin":{"usd":63872},"ethereum":{"usd":1793.11}}
Your LLM doesn't know the live price. One call, $0.01 — no monthly Pro plan, no API key.
paid-test record — how we bought it + on-chain receipt
Tested2026-06-04 BoughtGET pro-api.coingecko.com/api/v3/x402/simple/price · $0.01 Receipt0xd1e3…f50dae ↗ · Base · USDC
Messari ★ real brand · official x402$0.10
Institutional-grade asset metrics & multi-window ROI across the asset universe.
Endpoint
GET api.messari.io/metrics/v2/assets/roi?assets=bitcoin,ethereum
What came back · $0.10 · ~456 KB
{ "error": null, "data": [ { "name": "…", "symbol": "…", "returnOnInvestment": { "priceChange24h": …, "priceChange7d": …, "priceChange30d": …, "priceChangeYTD": … } }, … ] }
You can't reproduce institutional research data. $0.10 a pull vs a $25–99/mo seat — wins until ~100 queries/month.
paid-test record — how we bought it + on-chain receipt
Tested2026-05-20 BoughtGET api.messari.io/metrics/v2/assets/roi · $0.10 Receipt0x8156…65ba83 ↗ · Base · USDC
SignalFuse ⚠ on-demand signal$0.001
Trading signals (e.g. RSI reversion strategies).
Endpoint
GET api.signalfuse.co/v1/arena/rsi_reversion_swing/ETH
What came back · honest: no active signal at test time
{"strategy_id":"rsi_reversion_swing","symbol":"ETH","signal":null,"reason":"no_signal"}
Buy a signal per call instead of a black-box subscription. Its buyers are genuinely independent (we mis-flagged it once, then corrected — see how we test). Also offers free trial credits + a browser playground.
paid-test record — how we bought it + on-chain receipt
Tested2026-06-04 · re-tested BoughtGET api.signalfuse.co/v1/arena/rsi_reversion_swing/ETH · $0.001 Receipt0xcc86…1856c6 ↗ · Base · USDC

On-chain & DeFi — building this yourself is the expensive path

Alchemy ★ real brand · official x402$0.001
Ethereum / Base RPC node access (JSON-RPC).
Endpoint
POST x402.alchemy.com/eth-mainnet/v2 · body: {"jsonrpc":"2.0","method":"eth_blockNumber","params":[],"id":1}
What came back · $0.001
{"jsonrpc":"2.0","id":1,"result":"0x181262f"} // live ETH mainnet block
Running your own full node costs real money and ops time — the textbook "don't build this." $0.001 for one query.
paid-test record — how we bought it + on-chain receipt
Tested2026-05-20 BoughtPOST x402.alchemy.com/eth-mainnet/v2 · $0.001 Receipt0x3765…eeda5b ↗ · Base · USDC
OneSource$0.001
Live on-chain metadata — chain id, block height, gas price in one call.
Endpoint
GET skills.onesource.io/api/chain/network-info?network=base
What came back · $0.001
{ "data": { "block_number":"0x2cb4526", "chain_id":"0x2105", "gas_price":"0x5b8d80", "net_version":"8453", "network":"base" } }
Current chain state isn't something an LLM holds. $0.001 gets it live, no node, no setup.
paid-test record — how we bought it + on-chain receipt
Tested2026-05-20 BoughtGET skills.onesource.io/api/chain/network-info?network=base · $0.001 Receipt0x6c35…485a59 ↗ · Base · USDC
AgenticFi DeFi ⚠ covers CEX too$0.001
TVL & metrics by protocol — top 50 with chain breakdown. Covers CEX and DeFi.
Endpoint
GET defi-api.agenticfi.wtf/api/defi/protocol-tvl
What came back · $0.001 · ~12 KB, top entries
[ { "name":"Binance CEX", "tvl":141493809731, "category":"CEX", … }, { "name":"OKX", "tvl":22193109351, "category":"CEX", … }, … ]
Live TVL means reading many sources — a real build to maintain. One call, $0.001. Just note it leads with CEX entries.
paid-test record — how we bought it + on-chain receipt
Tested2026-06-04 · re-tested BoughtGET defi-api.agenticfi.wtf/api/defi/protocol-tvl · $0.001 Receipt0x5f61…aa1581 ↗ · Base · USDC

Real-world data — proprietary or live, you can't fake it

PayAPI Weather$0.001
Live weather by location.
Endpoint
POST weather.payapi.market/current · body: { "location": "London" }
What came back · $0.001 · live conditions
{ "location": { "name":"London", "country":"United Kingdom" }, "current": { "temperature_c":18.6, "feels_like_c":15.0, "humidity_pct":49, "wind_speed_kmh":20.9, "weather_description":"Partly cloudy", "time":"2026-06-04T19:15" } }
Free weather APIs exist — but they want a signup and a key your agent must hold. One call, no setup. (Note: the body field is location, not city — cost us a test to find out.)
paid-test record — how we bought it + on-chain receipt
Tested2026-06-04 · re-tested BoughtPOST weather.payapi.market/current · body {location:"London"} · $0.001 Receipt0xf998…5aec90 ↗ · Base · USDC
What we don't promise: that a service stays online (dormancy is common), that response content is always useful (SignalFuse honestly returned no signal), or that pricing won't change. Endpoints and responses are from real paid tests on 2026-06-04 — "worked when we checked," not a standing guarantee.
03 — What to avoid

What we paid for so you don't have to

We only flag a service here with on-chain, reproducible evidence — usually a payment we actually made.

lowpaymentfee.com — confirmed content fraud

Charges for a paid API and returns fabricated / worthless content. We have a full forensic write-up with on-chain transactions. The clearest "this is exactly what Atlas exists to catch" case.

Capminal — paid, got only an echo

Billed as an "AI-powered Token Deep Research API." We paid $0.01 with the correct token parameters and got back nothing but our own input echoed — no research, no data:

{"success":true,"message":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913"}
Pattern to distrust: "high volume = trustworthy"

In a market where most volume is wash, a high transaction count is not a quality signal — often the opposite. Don't pick services off a volume leaderboard. A pattern warning, not an accusation against any named service.

04 — Where you've landed

The market, measured — not guessed

We don't ask you to take "the market is mostly fake" on faith. Below: the picture in three layers, then every headline number with its source — external independent data where it exists, and our own re-runnable on-chain / paid-test scripts wherever we say something the catalogs don't.

On-chain data through 2026-05; ecosystem figures updated 2026-05-29. All scripts live in our repo and reproduce every cell.

The three layers

1 · The whole x402 ecosystem background

Coinbase-led HTTP-402 + USDC standard, now in the Linux Foundation. Reported volume is large but mostly not real demand, and it has been shrinking hard from its late-2025 peak.

~81% volume gamed (Artemis)
$5.15M → $1.19M monthly gross, Nov-25 → May-26
~$1.6M real / 30d (Artemis, wash-stripped)

2 · The agent marketplace (Bazaar) deep-dived

The largest agent-service catalog, and the one we audited end-to-end. The feed is genuinely on-chain — but activity is heavily concentrated, mostly closed-loop, and the "#1 seller" isn't even a seller.

~750K tx / $263K rolling gross
top-10 = 78.8% of volume · top-1 = 39%
single-digit%–19% is real organic

3 · Everything else partly unknown

The marketplace we deep-dived is only ~34% of the ecosystem. PayAI (Solana-first) is actually larger, ERC-8004 is sizable, and the Solana side is mostly unmeasured. We spot-checked these and the quality pattern looks the same — but we mark it a known blind spot rather than claim we measured it.

~43% PayAI
~34% Bazaar (Base)
~20% ERC-8004

How we measured it — every number, traceable

Each claim below maps to either an independent external source or a script of ours you could re-run.

Why ranking by volume will mislead you

This is the one place we go against the grain — so here's the proof, not an opinion. The usual catalogs rank services by on-chain volume. In this market that surfaces exactly the wrong things, and we can show it three ways:

That's why Atlas doesn't rank by volume. We pay to call a service and report what came back. Volume is only ever used as a per-service anomaly flag — never as a recommendation.

And one number we threw out: we once estimated wash with a buyer-concentration "ratio" and got ~98%. On scrutiny the ratio conflated heavy real usage with self-dealing, so we dropped it and rebuilt on closed-loop + paid-test evidence — which is why our defensible figure is the ~40% floor, not the flashier 98%. Sources here disagree by up to ~18× and the market moves weekly; we report ranges with method + cutoff attached, never a single precise figure we can't defend. (More on corrections in how we test.)
05 — How we test

How we test — and why you can check us

Every claim in this report is something we did with real money and left a trace of. Here's the method.

We publish our own mistakes

  • We once used a buyer-concentration "ratio" to judge wash trading. It conflated heavy real usage with self-dealing — so we dropped it for closed-loop and paid-test evidence.
  • We first flagged SignalFuse as fake traction. On a closer check its buyers were genuinely independent — a real service. We corrected the call; it's on the whitelist.
  • We listed services with "copy this command" before re-testing. When we actually re-ran them, an outdated client couldn't pay the new network format, and one "research" API turned out to just echo input. We fixed the client, dropped the dead ones, and kept only what paid through and returned real data.

Keeping the corrections visible is the point. An auditor that hides its errors is just another black box.

Want more than these 8? The full verified dataset — all 128 verified services, each with its payTo, on-chain tx + amount, the captured response, and the exact request that produced it (490+ paid tests went into it, every one on-chain) — is the paid /verified API, callable by your agent over x402, pure x402 with no API key. live now.