Seed Protocol is a shared data layer backed by Ethereum and Arweave. No databases to manage. No silos to escape. With versioning, provenance, and portability built into every record.
Seed Protocol coordinates permissionless network effects around the simplest possible data pattern. No opinions about authorship, no assumptions about sensemaking — just a foundation anyone can build on.
A unique, permanent identifier for any piece of data. The anchor point that persists across every version and context. Created once, referenced forever.
createSeed(schemaUid)An immutable snapshot linked to a Seed. Content evolves while its lineage remains intact. Every change is recorded, nothing is lost.
createVersion(seedUid)The meaning layer. Schemas, properties, and relationships that let queriers determine context. The reader decides the interpretation, not the author.
schema.models → propertiesWhether you want to build on the SDK, understand the protocol, or explore how Seed fits your domain — we have a starting point for you.
Install the SDK, define schemas with React hooks, and start publishing versioned data in minutes. TypeScript-first with full browser and Node support.
Explore the design principles, the Seed/Version/Metadata pattern, on-chain attestations via EAS, and permanent storage on Arweave.
Content provenance, tamper-proof sourcing, and portable publication history.
Explore →Own your drafts, versions, and publishing history across any platform.
Explore →Verifiable research data, reproducible datasets, and citation permanence.
Explore →Versioned data pipelines, schema evolution, and permanent dataset lineage.
Explore →Build on shared data without walled gardens. Your users own their network effects.
Explore →See the full picture of how Seed Protocol applies across domains.
Browse all →The path to network effects has always required extractive monopolies. Seed enables permissionless network effects where participants aren't behind anyone's API. Build your own, or use someone else's — without fear of being rugged.
Making content last requires deep humility about predicting the future. Seed makes the minimum assumptions about authorship and sensemaking — the querier of the data determines the meaning, not the author.
Have your own walled garden, then build bridges to others. Seed adds a semantic indexing layer so anyone can find what they're looking for — but no one's view of the data is the same.