Sato Hub

Datasets

Open, citable data on the onchain agent stack

7 datasets, CC-BY-4.0, no key and no signup. Attribution: data by satohub.ai. Every entry below states what a row is, how it was derived, how often it changes, and how to cite it with the date it was read. Scores and liveness are informational signals about how open, active and verifiable a project is — not safety, quality or returns grades.

Onchain Agents Index

Daily · 382 listings

The scored catalogue of what onchain AI agents are built from, rebuilt daily from public evidence.

JSONCSVNDJSON·Fields, coverage and citation

Onchain Agent Economy Measurements

Weekly · 243 measurements in the current week

Weekly on-chain measurements of agent registries, launchpads, payment rails and account infrastructure.

JSONCSVNDJSON·Fields, coverage and citation

Crypto-relevant agent skills from public registries, each with a static disclosure of what it does.

JSONCSVNDJSON·Fields, coverage and citation

Sato Score Inputs

Daily · 379 scored listings

Every input the Sato Score is computed from, per listing, so the score can be reproduced.

JSON (signed)·Fields, coverage and citation

What was added, updated and removed since a date — the delta companion to the index.

JSON·Fields, coverage and citation

Sato Hub Measured Numbers

Daily to weekly, per measurement · 7 measurement pages

The published measurements, each on its own dated page with the query that produced it.

HTML (Dataset JSON-LD per page)·Fields, coverage and citation

Sato Hub Answered Questions

Daily · 151 answered questions

Builder questions answered by a deterministic function over the live records, dated.

HTML (QAPage JSON-LD per page)·Fields, coverage and citation

How to cite

Cite the dataset page, not the file: the page states the licence, the method and the coverage window, and it stays at the same URL when a format is added. Include the date you read it — a figure from this site is true of the day it was read. The pattern:

Sato Hub. <dataset>. https://satohub.ai/datasets/<slug>. Accessed YYYY-MM-DD.

Worked example, in three formats:

Sato Hub. "Onchain Agents Index." Sato Hub, updated 2026-09-14, accessed 2026-09-14. https://satohub.ai/datasets/onchain-agents-indexSato Hub. (2026, September 14). Onchain Agents Index [Data set]. Sato Hub. Retrieved September 14, 2026, from https://satohub.ai/datasets/onchain-agents-index@misc{satohub-datasets-onchain-agents-index-2026-09-14,
  title        = {Onchain Agents Index},
  author       = {{Sato Hub}},
  year         = {2026},
  month        = {9},
  howpublished = {\url{https://satohub.ai/datasets/onchain-agents-index}},
  urldate      = {2026-09-14},
  note         = {Data set, last updated 2026-09-14. Licensed CC-BY-4.0.}
}

Each dataset page carries its own copy-paste block. Machine copy of this catalogue: /api/datasets.json. Licence terms: CC-BY-4.0.

Questions this page answers

Is there an open dataset of onchain AI agent tools?
Yes. The Onchain Agents Index is a free CC-BY-4.0 dataset of the projects onchain AI agents are built from — frameworks, MCP servers, wallets, payment rails, data tools and live agents — rebuilt daily from public evidence and served as JSON, CSV and NDJSON at satohub.ai/api/export/index.json. No key and no signup.
How do I cite Sato Hub data?
Each dataset page carries a plain-text, APA and BibTeX citation with both the date the data was refreshed and the date you read it. The short form is: Sato Hub. <dataset>. https://satohub.ai/datasets/<slug>. Accessed YYYY-MM-DD. A figure from this site is only true of the day it was read, so the date belongs in the citation.
What licence is the data under, and can I use it commercially?
CC-BY-4.0, including commercial use, with attribution: data by satohub.ai. That covers the catalog exports, the weekly agent-economy measurements, the skill disclosures and the score inputs.
How often is the data updated?
The directory index and the score inputs are rebuilt daily. The agent-economy measurements and the skill disclosures are collected weekly. Each dataset page states its own cadence and the date its series begins.
What is deliberately not in these datasets?
Score history, provenance, trust components and commercial fields stay on the site and behind the MCP server rather than in the free export. Nothing about who reads or queries Sato Hub is published anywhere: these are datasets about the onchain agent stack, not about this site's traffic.

Related: Export quickstart · MCP server · Score methodology · Transparency report · House rules