Not a directory. A builder with a directory underneath.
For most of its life, Sato Hub introduced itself as a directory: browse onchain agent frameworks, skills, MCPs, wallets, and payment rails, filtered by chain and category. That was true, and useful, and also the wrong lead. Nobody wakes up wanting to browse a directory. They wake up wanting to build an agent that trades, pays, or moves funds onchain — and the directory was always just the parts bin for that job.
So the front door changed. Sato is now positioned as the agent builder for crypto — describe the onchain agent you want, and Sato helps assemble the stack it takes to build it. The library — the mapped set of skills, MCPs, tools, wallets, payment rails, and frameworks — didn't disappear. It moved one layer back, from headline to engine.
Why the order matters
A prompt is not an agent. An agent that can actually do something onchain needs a wallet with real spend controls, a way to talk to a chain or a DEX, a payment rail if it's going to move value, and a runtime to hold it all together. That's a stack, not a sentence. Most "agent builder" tools hand you a blank prompt and leave the stack as an exercise for the reader.
Sato inverts that. The builder is the front door because the question builders actually have isn't "what exists" — it's "what do I need, and does it hold up." The library answers the first half. The rest of the platform — the Sato Score, verification, deploy specs — answers the second.
What's underneath the builder
The supply layer hasn't shrunk; it's just no longer the pitch:
- ▸The library — the directory of frameworks, skills, MCPs, wallets, payment rails, and venues a builder can draw on, organized by category and chain.
- ▸The Sato Score — a 0–100 read on how open, active, and verifiable a listed resource is. It's a transparency and liveness signal, evidence collected against maintenance, code, docs, and independent verification — explicitly not a safety, quality, or returns grade. A high score means the evidence is there to check, not that the project is safe to trust with funds.
- ▸Verification — badges that separate self-reported claims from ones Sato reproduced or reviewed. Most of the ecosystem's "verified" labels mean someone filled out a form. On Sato, it means an install was actually run or a claim was actually reviewed.
- ▸The marketplace — the next layer, not live yet. Once the supply and trust layers are solid, agents and tools built on top of them get a path to be listed, hired, or licensed.
What this doesn't mean
Worth saying plainly, because "builder" gets overloaded fast: Sato does not one-click deploy an agent for you, and it does not run or own the agent once you've built it. It maps the stack — the pieces, the evidence on each piece, and where to get them. The assembly and the running are still yours. That distinction is the whole point of keeping the Sato Score honest: a mapped stack is not a managed one.
What to watch
The near-term work is making the builder flow itself sharper — matching a described agent to the right stack faster, and surfacing the Sato Score and verification status inline as builders assemble, not as an afterthought on a separate page. Longer-term, the marketplace layer starts to matter once there's enough verified supply and enough built agents to justify it. Start at [satohub.ai](https://satohub.ai) if you want to see where the stack currently stands.