Dataset versions
A named dataset. Each version is fixed to one CID; the name points at the current version and keeps its history.
The host for named, versioned bytes — dataset versions and artifact releases. mithril.fund owns no storage itself; the catalog will be datoms in the same datom plane graph.kotoba.cloud queries.
Upload is not available yet. This page states where things stand. Where the bytes will be kept (the byte plane) is not decided yet, so there is no write path to show here. Writes are designed to come from the CLI and agents, not the browser (CLI first).
A named dataset. Each version is fixed to one CID; the name points at the current version and keeps its history.
A named, versioned object: name → current CID + version history. Folders are not the model.
The name → version → CID mapping is recorded as datoms in the datom plane graph.kotoba.cloud queries — not in a second database.
You can compute a CID on your own machine: the public library kotoba-lang/tech-ipfs-specs-unixfs chunks even large files as UnixFS client-side and gives the same CID kubo does (ipfs add --cid-version=1 --raw-leaves). tech-ipfs-specs-unixfs ↗
No prices, quotas or counts are shown here; numbers appear only once they are measured.