Your World Is Protected
Your world is stored in a Postgres database. The canonical adapter separates core from storage, so swapping the database does not change canonical hashes. Core computes all canonical records; the adapter caches them.
✗
Connection error
How data independence works
Canonical adapter principle: Postgres, Supabase, and Prisma do not generate canonical IDs. IDs are determined by sha256(canonicalize(hashInput)). Core computes canonical records; adapter stores them in Postgres. Even if the DB is wiped, complete recovery from the canonical ledger is possible.
Accord core (canonical identity)
placeCanonicalObject()
createCanonicalUsage()
executeWorldMutation()
exportCanonicalGraph()
importCanonicalGraph()
rebuildProjectionFromCanonicalGraph()
verifyCanonicalGraph()
→
Persistence adapter (serialization / cache)
Postgres（Supabase）
Prisma ORM
canonicalHash でインデックス
canonical レコードを生成しない
canonical フィールドを決定しない
Postgres
Main DB — stores canonical records
ADAPTER — not canonical identity
Supabase
Postgres host + connection pooler
ADAPTER — not canonical identity
Prisma ORM
Schema management + query layer
ADAPTER — not canonical identity
Recovery path: import → rebuild → verify → worldResult: same_world → DB fully replaceable
Postgres 接続状態
Checking connection state…
canonical 台帳エクスポート
Checking export state…
DB recovery steps
The DB is a cache layer. Even if wiped, complete recovery from the canonical ledger (accord-ledger.json) is possible.
1. export → download accord-ledger.json
2. POST /api/canonical/import
3. POST /api/canonical/rebuild
4. POST /api/canonical/verify → worldResult: same_world
5. All canonical hashes match — DB fully replaceable
Canonical Recovery