71 '<!-- CONFIRM -->' review markers across 19 files used HTML comment
syntax, which MDX cannot parse. One parse error breaks the whole
production build — this is why the deployed site 404'd on every page
while local dev limped along. All converted to {/* */} (code-fence
contents untouched). Also: remove the legacy source-'/' redirect,
replace the phantom architecture-diagram image with an ASCII diagram
until the real one lands.
Verified locally: mintlify broken-links parses all pages clean (one
known-good /api-reference tab link that 307s at runtime), and /,
/overview, /concepts/architecture, /quickstart, /patterns/*,
/versioning all render 200 with content.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
/ resolved into overview/use-cases because both overview.mdx and the
overview/ directory existed — move use-cases to /use-cases (redirected)
and remove the folder. Rewrite the landing as a mem0-style home: one
paragraph, the two-call loop, then card grids (start here / pick your
door / go deep / why) instead of the long-form pitch.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Serving the landing at / via an index page loops on Mintlify hosting
(/ -> first nav page /index -> normalized back to /). Use the /overview
slug instead: / redirects there once via Mintlify's built-in behavior,
/intro and legacy intro paths redirect to /overview, no destination is
bare / anymore.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Rename intro.mdx to index.mdx so Mintlify serves it at / directly; /intro
redirects to /, old redirect destinations follow, internal links updated.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The hosted spec at api.supermemory.ai/v3/openapi fails Mintlify's OpenAPI
validation: PATCH /brain/models 200 response is missing the required
'description' field, which aborted the whole docs deploy. Vendor the spec
into the repo with the missing description patched, and point both the api
config and the API Reference tab at the local file so deploys no longer
depend on fetching (or the validity of) the live endpoint.
Root-cause fix still needed API-side: add the description to the
/brain/models PATCH route's 200 response in the OpenAPI generator.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The monorepo's installed supermemory@3.10.0 lags npm. In the published SDKs
client.memories.add does not exist — adding is client.add() (top-level) or
client.documents.add(); client.memories has only forget/updateMemory. Swept
all 21 affected pages. Also: searchMode ('memories'|'hybrid'|'documents')
IS in the published typings and is now taught on hybrid-search as the mode
selector, with include.chunks marked deprecated back-compat; customId
charset corrected (dots, not colons); 14 CONFIRM markers resolved by the
published typings (entityContext, forget reason, Python signatures).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Developer Platform now reads: Start Here -> The Context Engine (concepts-
first: architecture, graph, hybrid search, profiles, permissioning,
surfaces) -> Using supermemory (+ versioning, errors-and-limits) ->
Building on supermemory (patterns) -> Connectors (+ faq, sync-lifecycle)
-> Ops and trust (analytics rejoins nav) -> Self-Hosting (+ tiers,
troubleshooting) -> Migration. container-tags/filtering fold into
permissioning/hybrid-search with redirects; 39 leftover files already
shadowed by redirects removed; every removed path redirects to its next
best page.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Fixes across existing pages: stale claude-3-sonnet -> current models,
deprecated ai/react + toAIStreamResponse -> current AI SDK APIs, five
divergent search signatures unified to client.search.memories (v4) /
client.search.documents (v3), singular containerTag in v4 contexts,
from-zep containerTag type error, canonical processing-status enum,
corrected 'v4 has no SDK' claim, internal links re-pointed at final
destinations.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
New concepts spine (architecture, hybrid-search, permissioning, surfaces,
glossary) built on one canonical mental model: ingest -> derive memories/
graph/profiles, one engine behind every surface. New Building-on-supermemory
pattern guides (multi-tenant, companion, multi-agent, task memory, company
brain, ingestion). New ops/trust pages (versioning, errors-and-limits,
usage-and-billing, security), connector FAQ + sync lifecycle from real
support answers, MCP + self-hosting troubleshooting, llms.txt for coding
agents. Every code sample verified against SDK types and backend routes;
unverified claims carry CONFIRM comments for review.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Every removed page gets a redirect to its next best live page. Salvage-kept:
analytics (only rate-limit/retention docs), integrations/cartesia (only
Cartesia SDK docs), memorybench/supported-models (now added to nav — live
pages link to it). Also removes stray test.py/test.ts/install.md from the
published docs directory.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The claude_code, opencode and cursor plugins are now free-tier
(supermemoryai/mono#2570), so the 'requires the Supermemory Pro plan'
warning on these integration pages is no longer accurate. OpenClaw
stays paid, so its banner is left in place.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds an admin-gated Models tab (main/triage/research pickers) that reads/writes the mono /brain/models endpoint, shown only for Company Brain orgs. Extracts a shared useOrgMemberRole hook so the brain settings sections dedupe the getActiveMember call.
Fixes ENG-1054
Co-authored-by: Dhravya Shah <dhravya@supermemory.com>
Co-authored-by: Dhravya Shah <dhravyashah@gmail.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Adds the **Proactiveness** settings tab for Company Brain channel/DM automations.
- Accordion list of automations — collapsed rows with an instant enable toggle + **Run now**, expand to edit.
- Two-column editor: prompt (left) / deliver-to channel or DM, frequency, day, time (right), with local-timezone-aware cron.
- Preset gallery (connection-first, category-diverse) for empty state + a New-automation menu.
- DM delivery option with a tooltip explaining channel visibility + personal-connection fallback.
- Profile-menu entry (gated on Company Brain).
Pairs with the API automations work: **supermemoryai/mono#2480**.
- Confirm-domain step, then live research transcript + action rail
- Poll research status; client-side force-start fallback
---
**Session Details**
- Session: [View Session](https://supermemory.us1.vorflux.com/agent-sessions/4f4c1321-9b65-4dac-9906-8b78c4c32926)
- Requested by: Unknown
- Address comments on this PR. Add `(aside)` to your comment to have me ignore it.
### TL;DR
Adds a swipeable "Review suggestions" card to the dashboard that lets users approve or decline inferred memories surfaced by Nova.
### What changed?
- Added a `ReviewMemoriesCard` component that appears in the "Suggested for you" section of the dashboard (both desktop and mobile layouts). The card is hidden when there are no pending inferred memories, so it never renders empty chrome. While the modal is open, the displayed count is frozen so the trigger button doesn't tick down or disappear mid-session. Switching spaces closes the modal automatically.
- Added a `ReviewMemoriesModal` component that presents inferred memories as a swipeable card deck. Users can approve (swipe right / ✓), decline (swipe left / ✗), or skip each memory. The modal includes:
- Drag-to-swipe with a full-card color wash (green for keep, red for decline) and verdict pills that intensify as the swipe threshold approaches
- Keyboard support: `→` to approve, `←` to decline, `↓` or `Space` to skip, and `Cmd/Ctrl+Z` to undo
- An undo button that steps back one card and reverts the server-side decision, using refs to avoid stale-state bugs during rapid interactions
- A progress dot indicator showing position in the queue alongside a numeric counter
- A "All caught up" completion state summarising how many memories were kept
- Reduced-motion support via `useReducedMotion`
- The card queue is snapshotted when the modal opens so cache updates from review mutations don't reshuffle the stack mid-session
- Added a `useInferredMemories` hook to fetch the pending review queue for a given container tag, and a `useReviewInferredMemory` mutation hook that calls the review endpoint. On success it removes the reviewed entry from the cached queue directly; on undo it invalidates the query to refetch the restored memory from the server.
- Registered two new API schema entries: `GET /container-tags/:containerTag/inferred` to fetch the pending queue and `POST /container-tags/:containerTag/inferred/:memoryId/review` to submit an approve, decline, or undo action.
### How to test?
1. Ensure there are inferred memories pending review for a container tag.
2. Open the dashboard — a "Review suggestions" card should appear in the "Suggested for you" section showing the count of pending memories.
3. Click the card to open the modal and swipe or use the buttons/keyboard to approve, decline, or skip memories.
4. Verify that approved and declined memories are removed from the queue after each decision and that the completion state appears once all cards are reviewed.
5. Use the undo button or `Cmd/Ctrl+Z` to step back through decisions and confirm the server-side state is reverted correctly.
6. Confirm the card does not render when there are zero pending inferred memories.
7. Switch spaces while the modal is open and confirm it closes without carrying state into the new space.
### Why make this change?
Nova infers memories on behalf of users but may not always be fully confident in them. This feature gives users a lightweight, low-friction way to review and curate those suggestions directly from the dashboard, improving the quality and trustworthiness of their memory store.
Hide the Getting Started checklist when all 3 onboarding steps complete, letting Recent memories go full-width. Swap the Setup 3/3 stat tile to a live Last updated timestamp once setup is done.
---
**Session Details**
- Session: [View Session](https://supermemory.us1.vorflux.com/agent-sessions/afbb6d10-e42e-4b40-aac9-90aba9532ec7)
- Requested by: Unknown
- Address comments on this PR. Add `(aside)` to your comment to have me ignore it.
Connections board on brain home: featured app tiles with OAuth/static connect against brain/mcp-connections, agent preview prompts, connector icon set, and reworked company-brain connections in settings.
---
**Session Details**
- Session: [View Session](https://supermemory.us1.vorflux.com/agent-sessions/e37671c7-3893-48d6-ad66-73200826c04d)
- Requested by: Unknown
- Address comments on this PR. Add `(aside)` to your comment to have me ignore it.
- onboarding: unified About step, mode-aware Sources, Slack-focused Flows step, connect feedback (toast + connected state), auto-draft company description from domain
- brain-home: Active members stat + invite, real OneDrive icon
- useConnectorAccess hook so company_brain unlocks pro-tier connectors across onboarding + integrations
- fix company-brain-connections crash on empty connections
---
**Session Details**
- Session: [View Session](https://supermemory.us1.vorflux.com/agent-sessions/64a15b29-0848-42d1-af4c-138c5a27136f)
- Requested by: Unknown
- Address comments on this PR. Add `(aside)` to your comment to have me ignore it.
## Docs: add Forget Matching endpoint + fix stale Forget Memory docs
### What this does
- **Adds docs for the new** **`POST /v4/memories/forget-matching`** **endpoint** — semantic/promptable mass-forget. Covers `dryRun` (preview), `threshold`/`maxForget` safety bounds, the request/response shape, and `forgetBatchId`.
- **Corrects the existing "Forget Memory" section** to match the actual implementation.
### ⚠️ No API surface changed
This PR is **docs-only**. The existing forget endpoint's behavior/contract is untouched — the previous docs were simply **wrong** and described a route that has never existed:
| | Old docs (incorrect) | Actual implementation (unchanged) |
| --- | --- | --- |
| Method + path | `POST /v4/memories/{id}/forget` | `DELETE /v4/memories` |
| Body | — | `{ id \| content, containerTag, reason? }` |
The handler (`forgetMemory` in `apps/api/src/routes/v4/memories/handlers.ts`) was not modified — this just makes the docs reflect reality.
### Also
- Small accuracy cleanups (response field descriptions, realistic example IDs).
### TL;DR
Adds documentation for the Memory Review endpoints and Profile Buckets feature.
### What changed?
**Memory Review (`memory-review.mdx`)**
- Added a new documentation page covering the two inferred memory review endpoints: `GET /v3/container-tags/{containerTag}/inferred` and `POST /v3/container-tags/{containerTag}/inferred/{memoryId}/review`.
- Documents the three review actions (`approve`, `decline`, `undo`) and how each affects search ranking and memory state (`isInference`, `isForgotten`, `reviewStatus`).
- Includes request/response examples in both `fetch` and cURL, a field reference table, error codes, and a collapsible React Query hooks example for building a review UI.
- Registered the new page in `docs.json` under the "Manage Content" group and linked to it from the Memory Operations next steps.
**Profile Buckets (`user-profiles.mdx`)**
- Added a "Profile Buckets" section explaining custom topical categories (`preferences`, `goals`, `work`, etc.) as a complement to `static`/`dynamic` profile sections.
- Documents the `include`, `buckets`, and `filters` query parameters on the profile endpoint.
- Covers the `GET /v4/profile/buckets` endpoint for listing configured bucket definitions, with request/response examples and a field reference.
- Explains the `[Recent]` / `[Summary]` label convention used in bucket and dynamic profile entries.
- Updated the `ProfileResponse` TypeScript interface to mark `static` and `dynamic` as optional and add the `buckets` field.
### How to test?
- Navigate to the docs site and confirm "Memory Review" appears in the sidebar under "Manage Content".
- Verify all code examples render correctly and tabs switch between `fetch` and cURL variants.
- Confirm the React Query accordion expands and displays the TypeScript snippet.
- Check that the Profile Buckets section renders inline within the User Profiles page, including the response JSON blocks and the tip/note callouts.
### Why make this change?
Inferred (derived) graph memories are down-weighted in search until reviewed, but there was no documentation explaining how to surface or act on them. Similarly, profile buckets were a shipped feature with no public-facing docs. These additions give developers the reference material needed to build review UIs and use topical bucket filtering in their integrations.
## Summary
- Parse enabled MCP OAuth API keys from the integrations page key list.
- Show MCP as connected in MCP integration cards, the Active filter count, featured CTA, and active connections rail.
- Reuse the existing MCP metadata signal used elsewhere in the app (`sm_source: "mcp"` or `sm_kind: "mcp_oauth_exchange"`).
<img width="1918" height="657" alt="image" src="https://github.com/user-attachments/assets/0dc4dace-9d47-4f05-8aad-55bd947dfe1c" />
## Summary
- Fix delete-organization dialog focus when switching orgs inside Settings.
- Restyle delete-organization modal to match the app modal theme and remove extra organization icons.
- Make the organization switcher list scrollable when many orgs exist.
- Send Create organization directly to onboarding instead of opening the create-org modal.
- Stop onboarding from completing when org creation fails, show an error toast, and return existing users to the dashboard.
## Summary
- update the Claude Code plugin install step to use `/plugin install supermemory`
- keep the marketplace add command pointing at `supermemoryai/claude-supermemory`
## Testing
- Not run; copy-only change
## Summary
- Add disconnect actions for org and personal GitHub/Linear connections on the Company Brain settings page
- Show connected Slack workspace name and an admin-only **Reconnect Slack** button (pairs with API admin gate in supermemoryai/mono#1912)
- Fix infinite loading skeleton when `/brain/connections` fails by falling back to empty state with a toast
## Test plan
- [ ] Open Settings → Company Brain connections as org admin
- [ ] Confirm Slack team name shows when workspace is connected
- [ ] Confirm **Reconnect Slack** is visible for admin/owner only
- [ ] Connect and disconnect GitHub/Linear for org (admin) and personal scopes
- [ ] Simulate failed connections fetch (e.g. offline) and confirm page renders instead of infinite skeleton
- Promote integrations from ?view=integrations to real /integrations and nested /integrations/[card] routes; the page body is shared via AppExperience and useViewMode is path-aware.
- Legacy ?view= URLs (and /settings/integrations) redirect to the new routes for back-compat; middleware/ensure-workspace allow the public routes.
- Add ?connect=<plugin|provider> deeplink that opens a card's connect modal instantly with a loading state (e.g. Hermes API key).
New /org/invite/[invitationId] page so consumer-org invitees can view and accept or decline invitations in the app instead of the console.
Fixes ENG-811
