BYOF’s core primitive is much narrower than the product pitch suggests: the published schema stores only an address and a list of followed addresses.
The project thesis is social-graph portability. The README repeatedly frames the problem as users losing their contacts when they leave closed platforms and web3 apps having to rebuild social context from scratch.
The implementation stack is explicitly offchain and DID-centric: Ceramic, 3ID Connect, Self.ID, and DID tooling are presented as the way to give users control without requiring gas for each follow action.
The create-model script is especially revealing because it shows the exact boundary of the primitive. BYOF is fundamentally a model-definition and storage experiment, with dashboard and API layers built around that small graph object.
The project is best treated as a historical prototype rather than an active ecosystem standard. The repo metadata shows a public repository created in late 2021 and last pushed in early 2022.
Specific source signals worth retaining
README: best source for the project thesis, intended dashboard/API split, and the explicit portability argument.
create-model.mjs: strongest source for how the Ceramic/Glaze model is actually constructed and how little data the base object contains.
BYOFGRPH.json: useful confirmation that the durable schema is minimal and centered on follow-list portability rather than richer social semantics.
Repo metadata: useful for placing the project in time and treating it as an early design experiment rather than a currently evolving standard.
Why this entry belongs in the corpus
It strengthens the identity / social-graph side of the corpus with a small but revealing historical prototype.
It provides a contrast case for Ethereum Follow Protocol and related systems by showing an offchain DID/Ceramic path to portability instead of NFT-owned or onchain-managed list objects.
The reusable insight is that many crypto-social systems reduce to a simple question: where does the portable graph live, who controls edits, and how much meaning is left to downstream clients versus encoded into the base registry.