docs: record that system-memory distribution is centralized, not Nostr

b79105bb7dfe · AtlantisPleb · 3d ago · parent b157b67139c5

docs: record that system-memory distribution is centralized, not Nostr

The owner's decision on 2026-08-25 drops the project-15 Nostr transport and
carries memories on the openagents.com server instead. Section 6 named the
relay as the one channel; it now names the decision, keeps the superseded
design visible rather than deleting it, and says which rule survives the
substitution — one channel, whatever it is. The sections section 6 constrains
are transport-agnostic and stand as written.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SoZMfWRSGnf6FZX2Ar9rQ2

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SoZMfWRSGnf6FZX2Ar9rQ2

Deploy story

What this commit did to the running system — joined from the forge receipt chain, the part a commit page elsewhere cannot show.

Not deployed through the forge lane

No push, promotion, build, or deploy receipt references this commit (receipts are scanned over a bounded recent window). Changes shipped by full node replacement carry their proof in the release gate receipt instead.

Changed files

modified docs/memory/2026-08-25-system-memory-spec.md

Diff

1 file changed, +22 -6


## 6. Transport

Distribution rides the engram sync seam and nothing else. The remaining
project-15 piece — the real Nostr transport behind `EngramTransport`, built
on `nostr-effect` — is the same transport system memories use. No second
channel, no separate poll, no HTTP side door: a system memory is an
`EngramEvent`, `publish` enqueues it, `drain` delivers it, `fetch` retrieves
it by filter.
**Superseded, 2026-08-25, by the owner's decision.** Distribution does not
ride a Nostr relay. The project-15 Nostr transport is dropped from the
roadmap, and the centralized cloud path carries system memories instead: the
openagents.com Phoenix/Postgres server that already owns accounts, threads,
thread events, and grants, with the CLI, web, and API as its clients. The
same decision governs user and learned memories (#51), and the build issue
that would have delivered relay distribution (#63) is closed as not planned.

What survives the substitution is the rule this section exists to state:
**one channel, whatever it is.** No second channel, no separate poll, no
side door. Read the rest of this section with the server sync in the place
the relay held — the sections it constrains (admission, privacy tiers,
recall caps, the knowledge-base line) are transport-agnostic and stand as
written. The paragraphs below are kept as the superseded design rather than
deleted, because the decisions on top of the seam still describe what the
cloud path has to provide.

The superseded design read: distribution rides the engram sync seam and
nothing else; the real Nostr transport behind `EngramTransport`, built on
`nostr-effect`, is the same transport system memories use; a system memory
is an `EngramEvent`, `publish` enqueues it, `drain` delivers it, `fetch`
retrieves it by filter.

Decisions on top of that seam:


233	233
234	234		`## 6. Transport`
235	235
236		-	`Distribution rides the engram sync seam and nothing else. The remaining`
237		-	project-15 piece — the real Nostr transport behind `EngramTransport`, built
238		-	on `nostr-effect` — is the same transport system memories use. No second
239		-	`channel, no separate poll, no HTTP side door: a system memory is an`
240		-	`EngramEvent`, `publish` enqueues it, `drain` delivers it, `fetch` retrieves
241		-	`it by filter.`
	236	+	`Superseded, 2026-08-25, by the owner's decision. Distribution does not`
	237	+	`ride a Nostr relay. The project-15 Nostr transport is dropped from the`
	238	+	`roadmap, and the centralized cloud path carries system memories instead: the`
	239	+	`openagents.com Phoenix/Postgres server that already owns accounts, threads,`
	240	+	`thread events, and grants, with the CLI, web, and API as its clients. The`
	241	+	`same decision governs user and learned memories (#51), and the build issue`
	242	+	`that would have delivered relay distribution (#63) is closed as not planned.`
	243	+
	244	+	`What survives the substitution is the rule this section exists to state:`
	245	+	`one channel, whatever it is. No second channel, no separate poll, no`
	246	+	`side door. Read the rest of this section with the server sync in the place`
	247	+	`the relay held — the sections it constrains (admission, privacy tiers,`
	248	+	`recall caps, the knowledge-base line) are transport-agnostic and stand as`
	249	+	`written. The paragraphs below are kept as the superseded design rather than`
	250	+	`deleted, because the decisions on top of the seam still describe what the`
	251	+	`cloud path has to provide.`
	252	+
	253	+	`The superseded design read: distribution rides the engram sync seam and`
	254	+	nothing else; the real Nostr transport behind `EngramTransport`, built on
	255	+	`nostr-effect`, is the same transport system memories use; a system memory
	256	+	is an `EngramEvent`, `publish` enqueues it, `drain` delivers it, `fetch`
	257	+	`retrieves it by filter.`
242	258
243	259		`Decisions on top of that seam:`
244	260