One primitive. Three fields. Every food interaction on earth.
Every food system in history has failed to interoperate because each one invented its own data model. FoodBlock compresses the entire domain into three fields. That compression is what makes universal adoption viable for the first time: a protocol rigorous enough for a multinational supply chain is also simple enough that a reasoning AI can write a valid block from a sentence of plain English. The market stall trader and the logistics operator speak the same grammar.
Who is in the system. Every participant (farmer, maker, venue, courier, inspector, agent) needs an identity that other blocks can reference. Without this, there is no authorship and no trust graph.
Where things happen. A field, a kitchen, a warehouse, a market stall. Location is not optional metadata, it is the ground truth that connects food to the land it came from.
What moves through the system. An ingredient, a product, a surplus batch. The thing that a provenance chain is ultimately about: what it is, where it started, how it changed.
How one thing becomes another. Wheat becomes flour becomes bread. Each step is a block that references what came before, making the chain of custody traversable from any endpoint.
How value moves between participants. An order, a payment, a delivery, a gift. The economic layer of the protocol: the record that makes trade auditable without a central ledger.
What one participant says about another. A certification, a review, a sensor reading, a trust signal. Observation is how the system learns what is true beyond what participants declare about themselves.
Every industry that coordinates at scale has a shared data standard underneath it. Banking has SWIFT. The web has HTTP. Healthcare has HL7. What each has in common is that the protocol is not owned by any participant, it is the neutral substrate on which all participants compete.
A shared format lets systems built by different organisations, in different countries, with different technologies, interoperate without asking permission. No central authority to route through. No platform that extracts rent for connecting buyer to seller. The grammar is public. The infrastructure belongs to everyone who uses it.
A block's identifier is derived deterministically from its content. Any modification changes the hash. Every record is tamper-evident by construction, not by access control.
Change a single character in a block and the hash changes entirely. There is no way to alter the historical record without producing a block with a different identity, which every peer in the network will reject.
Hashing requires no network connection. A fisherman at sea, a farmer in a valley with no signal: both can generate cryptographically valid, signed blocks and sync them the moment connectivity returns. The record is intact either way.
Two actors generating the same block independently produce the same hash. No central authority assigns IDs. No coordination is required. The identifier emerges from the content itself.
Every FoodBlock can carry refs that point to other blocks. Follow them backwards and you reconstruct the full history of anything in the system: the chain of custody, block by block, each one signed and tamper-evident, back to the origin. A loaf of bread points to the baking, which points to the flour, which points to the mill, which points to the certification that said the grain was what the farmer claimed it was.
Bread (substance.product)
<- Baking (transform)
<- Dough (transform)
<- Flour <- Mill <- Certification
<- Water <- Location
<- Starter
<- 14 days of fermentationA signed lie remains a lie. Cryptography establishes who said what and when. It cannot establish whether it is true. Trust must be inferred from the wider pattern of conduct recorded in the graph over time.
No single entity can own the standard through which an economy records its food data.
Autonomous agents are first-class actors in the FoodBlock protocol.