An autonomous agent can receive money. But can it verify what it's holding?
Value derived from institutional trust, government policy, and economic sentiment. None of this is independently verifiable by code. An agent consuming fiat data is trusting its context window.
Exchange rates, inflation data, economic reports. All inputs an agent can consume but never independently verify. The context window is the attack surface.
21 million cap. Enforced by math. Every block header is cryptographically linked. An agent can verify the chain with code—no trust required.
Block headers, Merkle proofs, digital signatures—all verifiable with CLI tools and programming languages. The agent can check the math itself.
An LLM doesn't need to trust. It needs to verify. Bitcoin is the only money where every claim can be checked with a CLI.
21 million coins. Not a policy decision—a consensus rule. The agent can read the code and verify.
Each header commits to the previous via SHA-256. The chain's integrity is math, not trust.
Prove a transaction exists without downloading the full blockchain. Lightweight verification for agents.
Every transaction is cryptographically signed. The agent can verify ownership with secp256k1.
bitcoin-cli, btcd, rust-bitcoin—dozens of implementations. The agent can pick its own tools.
No bank, no Fed, no API to trust. Just math and a peer-to-peer network the agent can query directly.
"Fiat currency requires faith. Bitcoin requires only computation. For an autonomous agent, the choice is obvious—trust what you can verify."
One CLI. Bitcoin + stablecoins. Built for machines.