Repository navigation
Say in the API that memory and Learning are kept per organization - #1697
Closed
davidmckayv wants to merge 3 commits into
Closed
davidmckayv wants to merge 3 commits into
davidmckayv wants to merge 3 commits into
Conversation
A person's memory and Learning spaces belong to one organization: nothing remembered or learned in one is listed, recalled or shared in another the person belongs to. The Learning and memory areas of the spec now say so, and that sharing a skill with every Bot means every Bot in the organization. Descriptions only; no operation or schema changes shape.
davidmckayv
force-pushed
the
orgs/learning-memory-api
branch
from
November 15, 2026 22:20
565d0c0 to
7362e46
Compare
…ver show its id Where OpenBot keeps a Learning container for each organization alone, the container's id is a digest that says nothing to a person. On the admin's Automatic learning page inside such an organization the container read "OpenBot: org-<digest> (org-<digest>)", the default's field held the digest, and the picker and each Bot's choice said "Deployment default". The API now says which container is the organization's own (organizationSpaceId on the Learning settings, null or absent where OpenBot keeps none), and the page chooses the words from that and the organization's name: - The container shown reads "Default container for <organization>", or 'The "<name>" container for <organization>' for one an admin named. - The picker says "This organization's default" and a Bot's choice "Use this organization's default". - A container field holds only the name an admin gave a container. An empty field is the organization's default. The page sends the id the server keeps. - A container that is not there yet, the button that creates it and the note about automatic assignment name it as the organization's default. Nothing that names a container is drawn until the server has said which one is the organization's own, so the id is never on the page. An install's first organization has no container kept for it alone, and reads as it always did: "Deployment default", the container's name and id, and the id in its field.
The Bot chat reconnects by itself when a run starts on its thread from another tab, another device or a routine. It learns of the run from a live update that names the thread, and takes it only when that name is its own thread id. A thread has two ids: the one the app keeps and sends, and the one live updates carry. In the organization default they are the same. In any other organization they differ, so an update about the open thread named an id the chat had never seen, and the answer appeared only after a reload. The server now says the live id when it mints a thread and when it is asked whether a thread is still there: liveThreadId, an optional addition to both conversation-threads answers. The Bot chat is given its live updates under the id it has, through a thread store registered for its Bot before the chat starts listening. Where the two ids are the same nothing is registered and the chat listens as it always has.
davidmckayv
added a commit
that referenced
this pull request
Nov 16, 2026
…ory and Learning, one's own organization and invitations, teams, and Desktop connected to a remote OpenBot (#1705) Organizations, phase 2, fused into one branch: each organization governs itself, memory and Learning are kept per organization, an organization of one's own at first sign-in with invitations, teams inside an organization, and OpenBot Desktop connected to an OpenBot that runs somewhere else. This branch merges, at their current heads: - #1701 Say in the spec and Admin that each organization governs itself - #1697 Say in the API that memory and Learning are kept per organization - #1700 Invitations, organization settings and an organization of one's own in the app - #1699 Teams inside an organization: the API, the screens, and granting to a team - #1698 Add an API keys page to Settings, and say what OAuth needs in the spec - #1702 Connect OpenBot Desktop to an OpenBot that runs somewhere else - main, with #1704 On top of the merges: - The spec bundled once and the clients generated again from it, keeping every piece's additions in the shared spec files. - The teams tests' organization and member fixtures carry the fields an organization and a member answer with (`personal`, `removable`, `assignableRoles`). - The rust-toolchain action's pin is commented with the tag it is (`v1`), as the static analysis of the workflows asks. Checked locally: format, lint, typecheck, API lint, the teams, Learning and audit tests. Draft: the pieces' own PRs stay open; this lands them together after one end-to-end test.
Contributor
Author
|
Carried in #1705. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
The Learning and memory areas of the API spec now say that both are kept per organization.
Why
An organization is a permissions boundary: a person can be in two, and sees each on its own.
Scope
Descriptions only. No operation, parameter or schema changes shape, so the breaking-change check has nothing to flag. The bundled spec and the generated clients are regenerated (
bun run api:generate); the only differences are the description text they carry.