Every figure below is the sentence the matching Sato Hub numbers page is serving right now, quoted as it stands, with the week-over-week change of the same series beside it. Nothing here is restated in different words, and nothing is carried over from a previous week: a series with one reading shows "no prior reading" rather than a change of zero. Readings are as of 2026-09-14.
Agents registered on ERC-8004, by chain
The ERC-8004 Identity Registry's agent-id counter, read per chain on 10 chains: BNB Chain 350,113, Base 86,940, Ethereum 50,801 and 7 more. Never summed across chains, as of 2026-09-14.
| What | This week | Change on last week | |---|---|---| | BNB Chain | 350,113 registered | +11,630 | | Base | 86,940 registered | +1,834 | | Ethereum | 50,801 registered | +74 | | Gnosis | 4,130 registered | +3 | | Avalanche | 1,819 registered | +4 | | Arbitrum | 1,499 registered | +27 | | Polygon | 670 registered | +33 | | Optimism | 542 registered | 0 |
Method, coverage and the chains or pairs behind this figure: [https://satohub.ai/numbers/erc-8004-agents-by-chain](https://satohub.ai/numbers/erc-8004-agents-by-chain). Measured against [ERC-8004 — Trustless Agents (the EIP)](https://eips.ethereum.org/EIPS/eip-8004), [ERC-8004 reference implementation](https://github.com/ChaosChain/trustless-agents-erc-ri).
Endpoints that declare x402 and actually answer
Of 67 endpoints declaring x402, Sato Hub reached 63 and 2 answered HTTP 402 with a payment payload a client could parse; no payment was sent, as of 2026-09-14.
| What | This week | Change on last week |
|---|---|---|
| Endpoints answering a parseable HTTP 402 | 2 | no prior reading — the probe is a point-in-time run, not a banked weekly series |
| x402 sellers registered (venue stage selling) | unknown | no prior reading — one weekly reading banked so far |
| x402 settlements observed (venue stage settling) | unknown | no prior reading — one weekly reading banked so far |
Method, coverage and the chains or pairs behind this figure: [https://satohub.ai/numbers/x402-endpoints-that-answer](https://satohub.ai/numbers/x402-endpoints-that-answer). Measured against [x402 — protocol specification](https://github.com/coinbase/x402), [RFC 9110 §15.5.3 — HTTP 402 Payment Required](https://www.rfc-editor.org/rfc/rfc9110#name-402-payment-required).
Listed crypto MCP servers that answered a handshake
Of 62 listed crypto MCP endpoints, Sato Hub probed 62 in its latest daily check and 39 completed the initialize handshake (62.9%); a handshake proves a server is up, nothing more, as of 2026-09-14.
| What | This week | Change on last week |
|---|---|---|
| Endpoints that answered initialize | 39 | no prior reading — a daily point-in-time check of other people's servers, not a banked weekly series |
Method, coverage and the chains or pairs behind this figure: [https://satohub.ai/numbers/mcp-servers-that-answer](https://satohub.ai/numbers/mcp-servers-that-answer). Measured against [Model Context Protocol — the initialize handshake](https://modelcontextprotocol.io/specification/2025-06-18/basic/lifecycle).
The Uniswap fee tier carrying the most observed volume
Sato Hub reads the Uniswap v3 pools of 7 agent-relevant pairs on 3 chains and names, per pair, the fee tier that carried the most swap volume we observed; the 0.05% tier leads 2, as of 2026-09-14.
| What | This week | Change on last week | |---|---|---| | cbbtc-weth on Base | 0.05% tier leads | no prior reading — the log walk is capped, so the covered window is 0.23 days, not a week | | usdc-usdt on Base | 0.01% tier leads | no prior reading — the log walk is capped, so the covered window is 0.23 days, not a week | | weth-usdc on Base | 0.3% tier leads | no prior reading — the log walk is capped, so the covered window is 0.23 days, not a week | | usdc-usdt on Ethereum | no readable pool | no prior reading — the log walk is capped, so the covered window is under a days, not a week | | wbtc-weth on Ethereum | no readable pool | no prior reading — the log walk is capped, so the covered window is under a days, not a week | | weth-usdc on Ethereum | no readable pool | no prior reading — the log walk is capped, so the covered window is under a days, not a week |
Method, coverage and the chains or pairs behind this figure: [https://satohub.ai/numbers/uniswap-fee-tiers-for-agent-pairs](https://satohub.ai/numbers/uniswap-fee-tiers-for-agent-pairs). Measured against [Uniswap — fee tiers and how swap fees accrue](https://docs.uniswap.org/concepts/protocol/fees), [Uniswap v3 core — the pool contract these readings come from](https://github.com/Uniswap/v3-core).
Score movers
Across 375 listings carrying a Sato Score on both days between 2026-09-07 and 2026-09-14, 8 rose and 8 fell, 8 crossed a tier boundary, and 8 were scored for the first time. Every reading is a daily snapshot; a day nobody measured is a gap and is never carried forward, and the Sato Score is how open, active and verifiable a project is — not a safety, quality or returns grade.
| What | This week | Change on last week | |---|---|---| | Q402 (Quack AI) | 67 (Medium) | +29 · maintenance 20→35 | | Hashnet MCP Server | 64 (Medium) | +15 · maintenance 20→35 | | Hedera MCP | 78 (High) | +15 · maintenance 20→35 | | LunarCrush MCP | 20 (Low) | -30 · maintenance 35→8 | | Solana LLM Oracle (SLO) | 31 (Low) | -15 · maintenance 20→5 | | create-solana-agent | 42 (Medium) | -15 · maintenance 20→5 |
Method, coverage and the chains or pairs behind this figure: [https://satohub.ai/movers](https://satohub.ai/movers). Measured against [Q402 (Quack AI) — the project's own source](https://github.com/quackai-org/q402-mcp), [Hashnet MCP Server — the project's own source](https://github.com/hashgraph-online/hashnet-mcp-js), [Hedera MCP — the project's own source](https://github.com/ExpertVagabond/hedera-mcp), [LunarCrush MCP — the project's own source](https://lunarcrush.com/products/lunarcrush-mcpserver).
What arrived in the directory
8 listings were added to the Sato Hub directory and 25 published a release in the last 7 days, as of 2026-09-14: hood.fun, Maestro (Telegram trading bot), minia2a, Ogena token risk, Pons, Pools.trade. Each is an observed event with a date and a link to its own source, never an endorsement.
The rows, with the source for each: [https://satohub.ai/launches](https://satohub.ai/launches).
What we corrected
Over the last 6 months Sato Hub added 257 listings and retired 1, recorded 0 verification changes, applied 0 corrections from 0 verified claims, and decided 3 submissions. The ledger behind that sentence, month by month, is at [https://satohub.ai/transparency](https://satohub.ai/transparency).
How to check any of this
Every figure above is derived at page-render time from Sato Hub's own tables and from public chain and registry reads. The free catalogue export is at [https://satohub.ai/api/export/index.json](https://satohub.ai/api/export/index.json) and the current-week adoption rows at [https://satohub.ai/api/export/adoption.json](https://satohub.ai/api/export/adoption.json), both CC-BY-4.0. Over MCP: onchain_agent_get_metrics, onchain_agent_get_trend and onchain_agent_explain_number, which returns the method for any of these.
A figure here is never summed across venues, chains or stages — those rows do not add — a sampled rate is never multiplied by a population, and an unknown is an unknown, never a zero.