Two founders who already built the trust layer once.

Fidacy is built by a team that spent a decade on exactly the terrain the product now automates. Together it built and scaled a regulated cross-border business to roughly R$25M in annual revenue, on the same questions Fidacy now answers for agents: who is this party, are they authorized, and can you prove it later. It exited that venture deliberately, on its own terms.

That range is why the product is both hard and complete. The company designs the difficult parts directly: the Know-Your-Agent identity layer, action mandates that can be issued, delegated and revoked, deterministic decisioning, EdDSA-signed verdicts anyone can verify against a public JWKS, a payment firewall that turns an approval into a short-lived grant before money moves, and an append-only audit trail whose head is anchored to Bitcoin. It runs in production today, rail-agnostic across AP2, A2A, cards, and stablecoins, and installs with a single MCP server. The same hands shipped production AI systems for years, from agents inside Oracle and SQL Server databases before that was common to Licitagram, an AI platform for public procurement now across Latin America and Europe.

Around that engineering sits the discipline financial and enterprise buyers demand: a banking background, an operator's command of structure and process, compliance and procurement readiness, and the commercial instinct for the relationships that move a deal. Multilingual and multi-jurisdictional, with a presence across the US, EU, and Brazil, the company is built to be trusted globally from day one. Deep engineering and regulated-industry rigor in one team is the moat, and what turns a thesis into infrastructure the agent economy can rely on.