Money that actually leaves
This is the honest P&L of the business as it runs today: external data and inference are the only costs that leave the building.
Brokro sells one thing: a counterparty dossier where every claim resolves to raw bytes a merchant actually returned. It is priced per call over x402, settled by your own agent’s signature, and charged only when deterministic verification passes end to end.
There is no plan to choose, no key to provision and no minimum. An agent discovers the price from /.well-known/x402 and pays inline.
Read live from third-party 402 headers at page load. These are independent businesses setting their own prices — we don’t control this column.
| Input | Counterparty | Live price | Chain | Real external cash? |
|---|---|---|---|---|
| Otto AI · Token Security | Otto AI (independent x402 merchant) | $0.0010 | eip155:8453 | yes |
| Otto AI · DEX Token Price | Otto AI (independent x402 merchant) | $0.0010 | eip155:8453 | yes |
| PrintMoneyLab · Kimchi Premium | PrintMoneyLab (independent x402 merchant) | $0.0020 | eip155:8453 | yes |
| IP402 · Geolocation & ASN | IP402 (independent x402 merchant) | free | — | no |
| Atlas DNS & Domain Posture Feed | Brokro own supply node | $0.0050 | eip155:2366 | no — internal transfer |
| Kite & Base EVM Contract Probe | Brokro own supply node | $0.0060 | eip155:2366 | no — internal transfer |
| Dolphin Open-Weight OSINT & Repo Telemetry | Brokro own supply node | $0.0060 | eip155:2366 | no — internal transfer |
| Sentinel Fallback Attestation Node | Brokro own supply node | $0.0050 | eip155:2366 | no — internal transfer |
| SERV Reasoning | screening · planning · cross-examination (3 model calls) | $0.0030–$0.0060 | off-chain | yes |
Most demos quote whichever number flatters them. Both are here because they answer different questions.
This is the honest P&L of the business as it runs today: external data and inference are the only costs that leave the building.
The conservative case: it shows the spread survives even if Brokro replaced its own DNS, EVM and OSINT feeds with paid third parties tomorrow.
The distinction matters. Brokro’s own supply nodes pay a treasury address Brokro controls. Counting that as cost would inflate the apparent cost base and counting it as revenue would inflate the apparent margin — so it is labelled an internal transfer and shown separately in both views.
A buyer’s agent that assembles four feeds itself still has to decide whether its own summary is faithful. Brokro moves that risk onto the seller: if the dossier doesn’t verify, the seller eats the cost of the work and the buyer is charged nothing.
Four merchants means four 402 negotiations, four chains to hold balances on, four failure modes and four prices that can move. Brokro collapses that into a single priced call on one chain, and absorbs a merchant going down or spiking its price.
Every call leaves a published receipt: what was bought, what was refused and why, the exact spend against the mandate, and a digest over all of it. That is what makes an agent’s purchase defensible after the fact.
Straight-line from the measured cash margin above. No assumed discounts, no assumed price increases.
| Verified calls / month | Revenue | External + inference cost | Net cash |
|---|---|---|---|
| 1,000 | $50 | $7–$10 | $40–$43 |
| 10,000 | $500 | $70–$100 | $400–$430 |
| 100,000 | $5,000 | $700–$1,000 | $4,000–$4,300 |
| 1,000,000 | $50,000 | $7,000–$10,000 | $40,000–$43,000 |
The cost line is dominated by inference, so it scales with calls rather than with headcount. The binding constraint on this business is third-party merchant availability, not margin.
Read from this deployment’s ledger, not a projection.
| Runs recorded | 0 |
| Paid wholesale calls | 0 |
| Wholesale revenue | $0.0000 |
| Upstream spend | $0.0000 |
| SERV inference spend | $0.0000 |
| Adversarial challenges blocked | 0 |
| External settlement mode | on |