## Summary
- Replaces per-provider TypeScript constant files
(`openai-models.const.ts`, `anthropic-models.const.ts`, etc.) with a
single `ai-providers.json` catalog as the source of truth
- Adds runtime model discovery via AI SDK for self-hosted providers,
with `models.dev` enrichment for pricing/capabilities
- Introduces composite model IDs (`provider/modelId`) for canonical,
conflict-free identification
- Simplifies provider configuration: API keys are injected from
environment variables (e.g., `OPENAI_API_KEY`)
- Adds admin panel UI for provider management (add/remove/test), model
discovery, recommended model configuration, and default fast/smart model
selection per workspace
- Removes deprecated config variables (`AI_DISABLED_MODEL_IDS`,
`AUTO_ENABLE_NEW_AI_MODELS`, etc.)
- Adds database migration for composite model ID format
## Test plan
- [ ] Server typecheck passes
- [ ] Frontend typecheck passes
- [ ] Server unit tests pass
- [ ] Frontend unit tests pass
- [ ] CI pipeline green
- [ ] Admin panel AI tab loads correctly
- [ ] Provider discovery works for configured providers
- [ ] Model recommendation toggles persist
- [ ] Default fast/smart model selection works
Made with [Cursor](https://cursor.com)
## Summary
Fixes#17941 — Saving a blank subdomain causes a redirect to
`.website.com`, effectively breaking the workspace.
**Root cause:** Three layers all fail to reject an empty string `""`:
1. **Frontend (`SettingsDomain.tsx`):** `SaveButton` has both
`onClick={onSave}` and `type="submit"`. The `onClick` fires first,
calling `handleSave()` directly without running Zod validation. So
`isDefined("")` returns `true`, the confirmation modal opens, and the
blank subdomain is submitted.
2. **Backend DTO (`update-workspace-input.ts`):** The `subdomain` field
has `@IsString()` + `@IsOptional()` but no pattern validation, so an
empty string passes the DTO layer.
3. **Backend service (`workspace.service.ts:152`):** `if
(payload.subdomain && ...)` — empty string is falsy in JS, so it skips
`validateSubdomainOrThrow()` entirely and writes `subdomain: ""` to the
database.
**The crash:** After save, the redirect logic does
`"myworkspace.website.com".replace("myworkspace", "")` →
`".website.com"`, sending the user to an invalid URL.
## Fix
- **Frontend:** Call `form.trigger()` at the start of `handleSave` to
run Zod validation regardless of whether the function was invoked via
`onClick` or `form.handleSubmit`. Returns early with validation error if
invalid.
- **Backend DTO:** Add `@Matches(/^[a-z0-9][a-z0-9-]{1,28}[a-z0-9]$/)`
to reject invalid subdomains at the request validation layer
(defense-in-depth).
- **Backend service:** Change `if (payload.subdomain && ...)` to `if
(isDefined(payload.subdomain) && ...)` so empty strings route through
`validateSubdomainOrThrow()` instead of being silently skipped.
## Test plan
- [x] Existing `is-subdomain-valid.util.spec.ts` tests pass (36/36)
- [x] TypeScript type checks pass for both `twenty-server` and
`twenty-front`
- [x] oxlint passes on all changed files
- [x] Prettier passes on all changed files
- [ ] Manual: Navigate to Settings > Domains, clear the subdomain field,
click Save — should show validation error, not redirect
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: Charles Bochet <charles@twenty.com>
Co-authored-by: Charles Bochet <charlesBochet@users.noreply.github.com>
## Summary
- Migrates `LogicFunctionModule`, `CodeInterpreterModule`, and
`CaptchaModule` from the `forRootAsync` + injection token pattern to the
`DriverFactoryBase` lazy-loading pattern (matching `EmailModule` and
`FileStorageModule`)
- Fixes#18724 where `LOGIC_FUNCTION_TYPE` was not respected in worker
processes because the driver was created at module boot time before the
DB config cache was loaded
- Removes `isEnvOnly` from `LOGIC_FUNCTION_TYPE`,
`CODE_INTERPRETER_TYPE`, `CAPTCHA_DRIVER`, `IS_MULTIWORKSPACE_ENABLED`,
and `FRONTEND_URL` — these can now be safely configured via the database
at runtime
## How it works
Each migrated module now uses a `DriverFactory` (extending
`DriverFactoryBase`) instead of a module-level async factory + Symbol
injection token:
1. **Lazy creation**: `getCurrentDriver()` creates the driver on first
call, after `DatabaseConfigDriver.onModuleInit()` has loaded the DB
cache
2. **Auto-recreation**: If config changes in the DB, the next
`getCurrentDriver()` call detects the key mismatch and creates a new
driver instance
3. **Unified config**: Both server and worker read from the same
database — driver config only needs to be set once
### Files deleted (old pattern)
- `logic-function-module.factory.ts`,
`logic-function-drivers.module.ts`, `logic-function-driver.constants.ts`
- `code-interpreter-module.factory.ts`
- `captcha.module-factory.ts`, `captcha-driver.constants.ts`
### Files created (new pattern)
- `logic-function-driver.factory.ts`
- `code-interpreter-driver.factory.ts`
- `captcha-driver.factory.ts`
Net: **-150 lines**
## Test plan
- [x] `npx nx typecheck twenty-server` passes
- [x] `npx nx lint:diff-with-main twenty-server` passes
- [ ] Integration tests pass (`npx nx run
twenty-server:test:integration:with-db-reset`)
- [ ] Verify logic functions execute in workflow runs (the original bug)
- [ ] Verify code interpreter works in workflow code steps
- [ ] Verify captcha validation works on sign-up (when captcha is
configured)
Made with [Cursor](https://cursor.com)
## Description
- Introduces a new engine command execution model that replaces the
previous approach of mapping `EngineComponentKey` to React components.
Instead, engine commands are now mounted headlessly via
`HeadlessEngineCommandMountRoot`, with their execution context populated
synchronously before mounting.
- Creates new headless command components
- Moves error handling from the SDK layer to the host app by wrapping
all mounted commands with a new `CommandMenuItemErrorBoundary`
The new flow works as follows:
- When a command menu item with an `engineComponentKey` is clicked,
`useCommandMenuItemFrontComponentCommands` calls
`useMountEngineCommand`, which synchronously reads the current context
store (object metadata, selected records, filters, view ID, etc.) and
writes a `MountedEngineCommandContext` into
`mountedEngineCommandsState`.
- The command is then mounted into `mountedEngineCommandsState`, which
triggers `HeadlessEngineCommandMountRoot` to render the corresponding
headless component from `ENGINE_COMPONENT_KEY_HEADLESS_COMPONENT_MAP`,
wrapped in `CommandMenuItemErrorBoundary`,
`ContextStoreComponentInstanceContext.Provider`, and
`EngineCommandComponentInstanceContext.Provider`.
- Each command component reads its execution context and delegates to
one of the 4 execution patterns: `HeadlessEngineCommandWrapperEffect`
(simple actions), `HeadlessConfirmationModalEngineCommandEffect`
(destructive actions needing confirmation),
`HeadlessNavigateEngineCommand` (GO_TO_* commands), or
`HeadlessOpenSidePanelPageEngineCommand` (SEARCH_RECORDS, ASK_AI,
VIEW_PREVIOUS_AI_CHATS).
- After execution, the command self-unmounts via
`useUnmountEngineCommand`, which removes the entry from
`mountedEngineCommandsState` and stops rendering the component.
### What
Unifies record page layout editing and navigation menu editing into a
single global "layout customization" session. Dashboard editing stays
separate.
### How it works
**Two edit mode systems, one context-based read:**
- `isLayoutCustomizationModeEnabledState` -- global atom for record
pages + navigation
- `isDashboardInEditModeComponentState` -- dashboard-only, independent
per-component atom
- `PageLayoutEditModeProvider` -- context that dispatches to
`RecordPageLayoutEditModeProvider` (reads global atom) or
`DashboardPageLayoutEditModeProvider` (reads component atom), one
component per file
**Session registry + independent atoms:**
- `activeCustomizationPageLayoutIdsState` -- accumulates page layout IDs
as user navigates during customization (`string[]`)
- Save/cancel iterate the ID list and read each layout's draft/persisted
atoms independently
- Follows the same pattern as `settingsRoleIdsState` +
`settingsDraftRoleFamilyState`
**Unified UI:**
- `LayoutCustomizationBar` replaces the old `NavigationMenuEditModeBar`
- Enter once -- edit record layouts + navigation -- save/cancel
everything together
- `useSaveLayoutCustomization` orchestrates sequential save: navigation
draft -- page layouts -- field widget groups
- Error snackbar on partial save failure (with TODO for future atomic
server mutation)
**Draft protection during customization:**
- `PageLayoutRelationWidgetsSyncEffect` guarded -- only updates
persisted state from server, skips draft/currentLayouts while
customization is active
- `useExecuteTasksOnAnyLocationChange` skips draft reset when
customization mode is enabled
- Command execution blocked during layout customization
### Cleanup
- Deleted `NavigationMenuEditModeBar`,
`isNavigationMenuInEditModeState`,
`isPageLayoutInEditModeComponentState`,
`useIsGlobalLayoutCustomizationActive`
- `DraftPageLayout` type changed from `Omit` to `Pick` (explicit fields)
- Removed save/cancel from `DefaultRecordCommandMenuItemsConfig` (bar
handles it now)
- Extracted `useSaveFieldsWidgetGroups` from save orchestration
- Split `PageLayoutEditModeProvider` into 3 separate files (one
component per file, Twenty convention)
### Known issues
- **Stale deleted widget after save (pre-existing on `main`)**: Delete
widget -- save -- exit customization -- Apollo cache stale -- sync
effect overwrites Jotai from stale data -- widget reappears until
refresh. Separate PR needed, likely tied to the planned server-side
`saveLayoutCustomization` atomic endpoint.
### Open questions
- **Module location**: Layout customization hooks/states live in `/app`
-- should they move to their own `modules/layout-customization/`?
- **Atomic server mutation**: All save mutations are on metadata schema
(`createNavigationMenuItem`, `deleteNavigationMenuItem`,
`updateNavigationMenuItem`, `updatePageLayoutWithTabsAndWidgets`,
`upsertFieldsWidget`). A single `saveLayoutCustomization` endpoint could
make saves truly atomic.
https://github.com/user-attachments/assets/036ef542-97f3-485b-a68f-3726002c81fb
## Summary
Uniformizes the metadata store to support **all** backend flat metadata
types, introduces a **minimal metadata endpoint** for fast initial
renders, replaces custom localStorage persistence with **Jotai's
built-in `atomWithStorage`**, and wires up a
**MinimalMetadataLoadEffect** for stale-while-revalidate loading.
### Key changes
- **All flat metadata types**: Added `FlatCommandMenuItem`,
`FlatFrontComponent`, `FlatWebhook`, `FlatRole`, `FlatRoleTarget`,
`FlatAgent`, `FlatSkill`, `FlatRowLevelPermissionPredicate`,
`FlatRowLevelPermissionPredicateGroup` — every entity in the backend
`MetadataEntityTypeMap` now has a corresponding frontend flat type
registered in `ALL_METADATA_ENTITY_KEYS` and `MetadataEntityTypeMap`.
- **Minimal metadata endpoint** (`minimalMetadata` GraphQL query): New
backend module (`MinimalMetadataModule`) returns lightweight object
metadata (names, icons, labels, flags) and basic views (id, type, key,
objectMetadataId) plus a `metadataVersion`. This enables fast first
paint before full metadata loads.
- **Jotai `atomWithStorage` for persistence**: Replaced the custom
`MetadataLocalStorageEffect` with Jotai's built-in `atomWithStorage` on
both `metadataStoreState` (family) and `metadataVersionState`. Added
`localStorageOptions` support to `createAtomFamilyState` for `{
getOnInit: true }` synchronous hydration. Each entity atom auto-persists
under keys like `metadataStoreState__objectMetadataItems`.
- **MinimalMetadataLoadEffect**: New effect mounted before
`MetadataProviderInitialEffects` that checks if the store already has
data (from Jotai localStorage hydration). If empty, it fetches minimal
metadata from the new endpoint. The full metadata load continues in
parallel, eventually enriching the store with complete data.
- **SSE effects alignment**: All metadata entity types now have
corresponding SSE effects that directly patch the metadata store via
`patchMetadataStoreFromSSEEvent`.
- **Existing selectors and joining logic**:
`objectMetadataItemsWithFieldsSelector`, `viewsWithRelationsSelector`,
`pageLayoutsWithRelationsSelector` reconstruct nested data from flat
entities for components that need it.
### Loading flow
```
App mount
→ Jotai atomWithStorage hydrates store from localStorage (sync, getOnInit)
→ MinimalMetadataLoadEffect
→ Store has data? → skip (app renders immediately)
→ Store empty? → fetch minimalMetadata endpoint → populate objects + views
→ MetadataProviderInitialEffects (full metadata load, runs in parallel)
→ LazyMetadataLoadEffect (page layouts, logic functions, nav menu, etc.)
→ IsAppMetadataReadyEffect (sets isAppMetadataReady)
```
## Test plan
- [ ] Verify app loads with empty localStorage (should fetch minimal
metadata, then full)
- [ ] Verify app loads with populated localStorage (should skip minimal
fetch, render immediately)
- [ ] Verify SSE events correctly update metadata store for all entity
types
- [ ] Verify logout clears metadata store (atom reset propagates to
localStorage)
- [ ] Verify all metadata selectors return correct joined data
- [ ] CI: lint, typecheck, tests pass
## Summary
- **Removes the entire `modules/favorites/` directory** (~66 files,
~5000 lines deleted) — components, hooks, states, types, utils, tests,
and the favorite-folder-picker sub-module
- **Eliminates the dual-write pattern** where creating a favorite also
created a NavigationMenuItem — all consumers now use
`useCreateNavigationMenuItem` directly
- **Removes `IS_NAVIGATION_MENU_ITEM_EDITING_ENABLED` feature flag
checks** from ~12 files, always taking the NavigationMenuItem code path
- **Cleans up backend dual-writes** in `object-metadata.service.ts` and
`twenty-standard-application.service.ts` that were creating Favorite
records alongside NavigationMenuItems
- **Updates prefetch system** to only load NavigationMenuItems (removes
favorites prefetch effects and states)
- **Cleans up test infrastructure** — updates Storybook decorators, mock
data, and graphql mocks to remove favorites references
### What was intentionally kept
- **Backend entity definitions** (`FavoriteWorkspaceEntity`,
`FavoriteFolderWorkspaceEntity`) — these define the database schema and
need a proper database migration to remove
- **Cascade deletion listeners** — still needed to clean up existing
Favorite data in workspaces that haven't been fully migrated
- **v1.18 migration commands** — needed for workspaces upgrading from
older versions
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
## Summary
Implements enterprise licensing and per-seat billing for self-hosted
environments, with Stripe as the single source of truth for subscription
data.
### Components
- **twenty-website** hosts the private key to sign `ENTERPRISE_KEY` and
`ENTERPRISE_VALIDITY_TOKEN`. It communicates with Stripe to emit the
daily `ENTERPRISE_VALIDITY_TOKEN` if the subscription is active, based
on the user's Stripe subscription ID stored in `ENTERPRISE_KEY`.
- **Stripe** is the single source of truth for subscription data
(status, seats, billing).
- **The client** (twenty-server + DB + workers) saves `ENTERPRISE_KEY`
in the `keyValuePair` table (or `.env` if
`IS_CONFIG_VARIABLES_IN_DB_ENABLED` is false) and the daily-renewed
`ENTERPRISE_VALIDITY_TOKEN` in the `appToken` table.
`ENTERPRISE_VALIDITY_TOKEN` is verified client-side using a public key
to grant access to enterprise features (RLS, SSO, audit logs, etc.).
### Flow
1. When requesting an upgrade to an enterprise plan (from **Enterprise**
in settings), the user is shown a modal to choose monthly/yearly
billing, then redirected to Stripe to enter payment details. After
checkout, they land on twenty-website where they are exposed to their
`ENTERPRISE_KEY`, which they paste in the UI. It is saved in the
`keyValuePair` table. On activation, a first `ENTERPRISE_VALIDITY_TOKEN`
with 30-day validity is stored in the `appToken` table.
2. **Every day**, a cron job runs and does two things:
- **Refreshes the validity token**: communicates with twenty-website to
get a new `ENTERPRISE_VALIDITY_TOKEN` with 30-day validity if the Stripe
subscription is still active. If the subscription is in cancellation,
the emitted token has a validity equal to the cancellation date. If it's
no longer valid, the token is not replaced. The cron only needs to run
every 30 days in practice, but runs daily so it's resilient to
occasional failures.
- **Reports seat count**: counts active (non-soft-deleted)
`UserWorkspace` entries and sends the count to twenty-website, which
updates the Stripe subscription quantity with proration. Seats are also
reported on first activation. If the subscription is canceled or
scheduled for cancellation, the seat update is skipped.
3. `ENTERPRISE_VALIDITY_TOKEN` is verified server-side via a public key
to grant access to enterprise features.
### Key concepts
Three distinct checks are exposed as GraphQL fields on `Workspace`:
| Field | Meaning |
|---|---|
| `hasValidEnterpriseKey` | Has any valid enterprise key (signed JWT
**or** legacy plain string) |
| `hasValidSignedEnterpriseKey` | `ENTERPRISE_KEY` is a properly signed
JWT (billing portal makes sense) |
| `hasValidEnterpriseValidityToken` | `ENTERPRISE_VALIDITY_TOKEN` is
present and not expired (expiration depends on signed token payload, not
on "expiresAt" on appToken table which is only indicative) |
Feature access is gated by `isValid()` =
`hasValidEnterpriseValidityToken || hasValidEnterpriseKey` (to support
both new and legacy keys during transition). After transition isValid()
= hasValidEnterpriseValidityToken
### Frontend states
The Enterprise settings page handles multiple states:
- **No key**: show "Get Enterprise" with checkout modal
- **Orphaned validity token** (token valid but no signed key): prompt
user to set a valid enterprise key
- **Active/trialing but no validity token**: show subscription status
with a "Reload validity token" action
- **Active/trialing**: show full subscription info, billing portal
access, cancel option
- **Cancellation scheduled**: show cancellation date, billing portal
- **Canceled**: show billing history link and option to start a new
subscription
- **Past due / Incomplete**: prompt to update payment or restart
### Temporary retro-compatibility: legacy plain-text keys
Previously, enterprise features were gated by a simple check: any
non-empty string in `ENTERPRISE_KEY` granted access. With this PR, we
transition to a controlled system relying on signed JWTs.
To avoid breaking existing self-hosted users:
- **Legacy plain-text keys still grant access** to enterprise features.
`hasValidEnterpriseKey` returns `true` for both signed JWTs and plain
strings, and `isValid()` checks `hasValidEnterpriseKey` as a fallback
when no validity token is present.
- **A deprecation banner** is shown at the top of the app when
`hasValidEnterpriseKey` is `true` but `hasValidSignedEnterpriseKey` is
`false`, informing the user that their key format is deprecated and they
should activate a new signed key.
- **No billing portal or subscription management** is available for
legacy keys since there is no Stripe subscription to manage.
This retro-compatibility will be removed in a future version. At that
point, `isValid()` will only check `hasValidEnterpriseValidityToken`.
### Edge cases
- **Air-gapped / production environments**: for self-hosted clients that
block external traffic (or for our own production), provide a long-lived
`ENTERPRISE_VALIDITY_TOKEN` (e.g. 99 years) directly in the `appToken`
table, with no `ENTERPRISE_KEY`. The daily cron will skip the refresh
(no enterprise key to authenticate with), but the pre-seeded validity
token will be used to grant feature access. No billing or seat reporting
occurs in this mode.
- **`IS_CONFIG_VARIABLES_IN_DB_ENABLED` is false**: if the user tries to
activate an enterprise key but DB config writes are disabled, the
backend returns a clear error asking them to add `ENTERPRISE_KEY` to
their `.env` file manually.
- **Canceled subscriptions**: the `/seats` endpoint skips Stripe updates
for canceled or cancellation-scheduled subscriptions to avoid Stripe API
errors.
### How to test
- launch twenty-website on a different url (eg localhost:1002)
- add ENTERPRISE_API_URL=http://localhost:3002/api/enterprise (or else)
in your server .env
- ask me for twenty-website's .env file content (STRIPE_SECRET_KEY;
STRIPE_ENTERPRISE_MONTHLY_PRICE_ID;STRIPE_ENTERPRISE_YEARLY_PRICE_ID;
ENTERPRISE_JWT_PRIVATE_KEY; ENTERPRISE_JWT_PUBLIC_KEY;
NEXT_PUBLIC_WEBSITE_URL)
- visit Admin panel / enterprise
actions are being renamed to command menu item, they will be migrated to
server and will be served as headless front components
---------
Co-authored-by: Raphaël Bosi <71827178+bosiraphael@users.noreply.github.com>
## Summary
- **Merge queue optimization**: Created a dedicated
`ci-merge-queue.yaml` workflow that only runs Playwright E2E tests on
`ubuntu-latest-8-cores`. Removed `merge_group` trigger from all 7
existing CI workflows (front, server, shared, website, sdk, zapier,
docker-compose). The merge queue goes from ~30+ parallel jobs to a
single focused E2E job.
- **Label-based merge queue simulation**: Added `run-merge-queue` label
support so developers can trigger the exact merge queue E2E pipeline on
any open PR before it enters the queue.
- **Prettier in lint**: Chained `prettier --check` into `lint` and
`prettier --write` into `lint --configuration=fix` across `nx.json`
defaults, `twenty-front`, and `twenty-server`. Prettier formatting
errors are now caught by `lint` and fixed by `lint:fix` /
`lint:diff-with-main --configuration=fix`.
## After merge (manual repo settings)
Update GitHub branch protection required status checks:
1. Remove old per-workflow merge queue checks (`ci-front-status-check`,
`ci-e2e-status-check`, `ci-server-status-check`, etc.)
2. Add `ci-merge-queue-status-check` as the required check for the merge
queue
## Summary
Front Before:
<img width="1199" height="670" alt="image"
src="https://github.com/user-attachments/assets/b978f67c-c0a6-49fc-bedd-a443f11c365d"
/>
Front After:
<img width="1199" height="670" alt="image"
src="https://github.com/user-attachments/assets/a4939dbb-a8b4-4c74-978c-daa7f27d00f3"
/>
Server Before:
<img width="1199" height="670" alt="image"
src="https://github.com/user-attachments/assets/da53e97f-ec65-4224-a656-ca41040aef6e"
/>
Server After:
<img width="1199" height="670" alt="image"
src="https://github.com/user-attachments/assets/8cdf3885-f515-4d6c-989f-a421a4e8206c"
/>
### CI Server Pipeline Restructuring
- Split monolithic `server-setup` job into three parallel jobs:
`server-build`, `server-lint-typecheck`, and `server-validation`
- `server-build` only handles build + Nx cache save (~1m vs old 3.5m),
unblocking downstream jobs faster
- `server-lint-typecheck` runs in parallel with no DB dependency
- `server-validation` handles DB setup, migration checks, and GraphQL
generation checks in parallel with tests
- Make `server-test` (unit tests) fully independent — no longer waits
for server-setup, builds its own artifacts
- Increase integration test shards from 8 to 10 for better parallelism
- Expected critical path reduction: ~10m → ~7m (~30% faster)
### CI Front Pipeline Improvements
- Use artifact upload/download for storybook build instead of rebuilding
in test shards
- Serve pre-built storybook via `http-server` in test jobs, with
`STORYBOOK_URL` env var
- Update `vitest.config.ts` to use `storybookUrl` when `STORYBOOK_URL`
is set
- Remove redundant `twenty-shared`, `twenty-ui`, `twenty-sdk` builds
from storybook test shards
### Vite Build Optimizations
- Conditionally enable `rollup-plugin-visualizer` behind `ANALYZE=true`
env var (not loaded by default)
- Broaden Istanbul coverage exclusions to skip test files, stories,
mocks, and decorators
- Remove `@tabler/icons-react` alias from twenty-front and storybook
configs
- Bundle `@tabler/icons-react` into twenty-ui instead of treating it as
an external dependency
- Add lazy loading with `React.lazy` + `Suspense` for all page-level
route components in `useCreateAppRouter`
This PR adds the necessary tool to create a demo workspace with :
relevant custom objects and fields, mock data and a real dashboard with
graph widgets.
It is still a bit under-optimized and slow but it works.
This PR also adds an AI tool that allows to see what happens in real
time, it navigates the app and waits when necessary.
---------
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
## Summary
- Implement a return-to-path mechanism that preserves the user's
intended destination across authentication flows (login, magic link,
cross-domain redirects)
- Uses layered persistence: Jotai atom (in-memory), sessionStorage with
TTL (tab-switch resilience), URL query parameter (cross-domain
propagation)
- Includes path validation to prevent open redirects, automatic cleanup
after successful login, and comprehensive test coverage
- Replaces the unused `previousUrlState` with a robust
`returnToPathState` system
## Test plan
- [ ] Visit a deep link (e.g. `/objects/tasks`) while logged out —
should redirect to login, then back to `/objects/tasks` after logging in
- [ ] Visit an OAuth authorize link while logged out — should redirect
to login, then to the authorize page
- [ ] Test magic link flow: click sign-in link that opens new tab —
should still redirect to original destination
- [ ] Test cross-domain: visit `app.twenty.com/objects/tasks` — should
preserve path through workspace domain redirect
- [ ] Verify auth/onboarding paths are excluded from being saved as
return paths
- [ ] Verify return-to-path is cleared after successful navigation
- [ ] All 215 existing `usePageChangeEffectNavigateLocation` tests pass
Made with [Cursor](https://cursor.com)
## Summary
Removes the `recoil` dependency entirely from `package.json` and
`twenty-front/package.json`, completing the migration to Jotai as the
sole state management library.
Removes all Recoil infrastructure: `RecoilRoot` wrapper from `App.tsx`
and test decorators, `RecoilDebugObserver`, Recoil-specific ESLint rules
(`use-getLoadable-and-getValue-to-get-atoms`,
`useRecoilCallback-has-dependency-array`), and legacy Recoil utility
hooks/types (`useRecoilComponentState`, `useRecoilComponentValue`,
`createComponentState`, `createFamilyState`, `getSnapshotValue`,
`cookieStorageEffect`, `localStorageEffect`, etc.).
Renames all `V2`-suffixed Jotai state files and types to their canonical
names (e.g., `ComponentStateV2` -> `ComponentState`,
`agentChatInputStateV2` -> `agentChatInputState`, `SelectorCallbacksV2`
-> `SelectorCallbacks`), and removes the now-redundant V1 counterparts.
Updates ~433 files across the codebase to use the renamed Jotai imports,
remove Recoil imports, and clean up test wrappers (`RecoilRootDecorator`
-> `JotaiRootDecorator`).
## Summary
- **Refactor frontend metadata loading architecture**: Split the
monolithic `EagerMetadataLoadEffect` into focused provider effects
(`UserMetadataProviderEffect`, `ObjectMetadataProviderEffect`,
`ViewMetadataProviderEffect`) orchestrated by `MetadataProviderEffects`.
Replaced `UserProvider` + `ObjectMetadataItemsProvider` with a single
`MetadataGater` that gates rendering on `isAppMetadataReadyState`. The
metadata store now validates view-object consistency before promoting
views, and `updateDraft` skips no-op updates via deep equality checks.
- **SDK CLI improvements**: Added `app:typecheck` command, improved
error handling in API sync (extracts GraphQL error messages), added
`serializeError` utility for human-readable error output, added `error`
file status to dev mode orchestrator with UI support, and fixed
ClickHouse migration/seed commands to use `transpile-only`.
## Description
- Add `isHeadless` field to `FrontComponent` entity so front components
can run without rendering UI in the command menu
- Introduce headless front component mounting logic:
`HeadlessFrontComponentMountRoot` at the application root,
`useMountHeadlessFrontComponent`, and `useUnmountHeadlessFrontComponent`
hooks to mount/unmount headless components
- Expand the SDK with new action components (`Action`, `ActionLink`,
`ActionOpenSidePanelPage`) and host communication functions
(`openSidePanelPage`, `unmountFrontComponent`)
- Move `CommandMenuPages` type to twenty-shared so the SDK can reference
it for side panel navigation
## Video QA
https://github.com/user-attachments/assets/4f9e3bb1-fcd1-42be-b3f4-a97e80c2add2
## Summary
- **Fix token renewal endpoint**: Use `/metadata` instead of `/graphql`
for token renewal in agent chat, fixing auth issues
- **Improve tool display**: Add `load_skills` support, show formatted
tool names (underscores → spaces) with finish/loading states, display
tool icons during loading, and support custom loading messages from tool
input
- **Refactor workflow agent management**: Replace direct
`AgentRepository` access with `AgentService` for create/delete/find
operations in workflow steps, improving encapsulation and consistency
- **Simplify Apollo client usage**: Remove explicit Apollo client
override in `useGetToolIndex`, add `AgentChatProvider` to
`AppRouterProviders`
- **Fix load-skill tool**: Change parameter type from `string` to `json`
for proper schema parsing
- **Update agent-chat-streaming**: Use `AgentService` for agent
resolution and tool registration instead of direct repository queries
## Test plan
- [ ] Verify AI agent chat works end-to-end (send message, receive
response)
- [ ] Verify tool steps display correctly with icons and proper messages
during loading and after completion
- [ ] Verify workflow AI agent step creation and deletion works
correctly
- [ ] Verify workflow version cloning preserves agent configuration
- [ ] Verify token renewal works when tokens expire during agent chat
Made with [Cursor](https://cursor.com)
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: cubic-dev-ai[bot] <191113872+cubic-dev-ai[bot]@users.noreply.github.com>
## Remove Recoil from twenty-ui
Completely removes the `recoil` dependency from `twenty-ui` by
converting all atoms, hooks, and providers to Jotai equivalents.
### twenty-ui
- `createState` now returns a Jotai `PrimitiveAtom` instead of a Recoil
atom
- `iconsState`, `IconsProvider`, `useIcons` converted to Jotai
(`useSetAtom`, `useAtomValue`)
- `RecoilRootDecorator` now uses Jotai `Provider` (name kept for compat)
- Deleted unused `invalidAvatarUrlsState` (Avatar already uses
`invalidAvatarUrlsAtomV2`)
- Removed `recoil` from `package.json`
### twenty-front
- Created local Recoil `createState` at
`@/ui/utilities/state/utils/createState` for ~112 state files still on
Recoil
- Updated all imports accordingly
- Removed `iconsState` from Recoil snapshot preservation in `useAuth`
(lives in Jotai store now)
## Recoil → Jotai progressive migration: infrastructure +
ChipFieldDisplay
### Benchmark
In the beginning, there was no hope:
<img width="1180" height="948" alt="image"
src="https://github.com/user-attachments/assets/f8635991-52e6-4958-8240-6ba7214132b2"
/>
Then the hope was reborn
<img width="2070" height="948" alt="image"
src="https://github.com/user-attachments/assets/be1182b9-1c8d-4fdc-ab4c-1484ad74449d"
/>
### Approach
We introduce a **V2 state management layer** backed by Jotai that
mirrors the existing Recoil API, enabling component-by-component
migration without a big-bang rewrite.
#### V2 API (Jotai-backed, Recoil-ergonomic)
- `createStateV2` / `createFamilyStateV2` — drop-in replacements for
`createState` / `createFamilyState`, returning wrapper types over Jotai
atoms
- `useRecoilValueV2`, `useRecoilStateV2`, `useFamilyRecoilValueV2`, etc.
— thin wrappers around Jotai's `useAtomValue` / `useAtom` / `useSetAtom`
- A shared `jotaiStore` (via `createStore()`) passed to a
`<JotaiProvider>` wrapping `<RecoilRoot>`, also accessible imperatively
for dual-writes
#### Dual-write bridge for progressive migration
For state shared between migrated and non-migrated components, we use
**dual-write**: writers update both the Recoil atom and the Jotai V2
atom (via `jotaiStore.set()`). This avoids sync components or extra
subscriptions.
Write sites updated: `useUpsertRecordsInStore`, `useSetRecordTableData`,
`ListenRecordUpdatesEffect`, `RecordShowEffect`,
`useLoadRecordIndexStates`, `useUpdateObjectViewOptions`.
#### First migration: ChipFieldDisplay render path
- `useChipFieldDisplay` → reads `recordStoreFamilyStateV2` via
`useFamilyRecoilValueV2` (was `useRecoilValue(recordStoreFamilyState)`)
- `RecordChip` → reads `recordIndexOpenRecordInStateV2` via
`useRecoilValueV2` (was `useRecoilValue(recordIndexOpenRecordInState)`)
- `Avatar` (twenty-ui) and event handlers (`useOpenRecordInCommandMenu`)
left on Recoil — not on the render path / in a different package
#### Pattern for migrating additional state
1. Create V2 atom: `createStateV2` or `createFamilyStateV2`
2. Add `jotaiStore.set(v2Atom, value)` at each write site
3. Switch readers to `useRecoilValueV2(v2Atom)`
4. Once all readers are migrated, remove the Recoil atom and dual-writes
#### Why not jotai-recoil-adapter?
Evaluated
[jotai-recoil-adapter](https://github.com/clockelliptic/jotai-recoil-adapter)
— not production-ready (21 open issues, no React 19, forces providerless
mode, missing types). We built a purpose-built thin layer instead.
## Reduce type leakage between GraphQL schemas
### Why
Twenty runs two separate GraphQL schemas: **core** and **metadata**.
NestJS's `@nestjs/graphql` uses a global `TypeMetadataStorage` that
accumulates all decorated types across all modules. When each schema is
built, every registered type leaks into both schemas regardless of which
module it belongs to.
This means the core schema's generated TypeScript
(`generated/graphql.ts`) contained ~2,700 lines of types that only
belong to the metadata schema (and vice versa). This creates confusion
about type ownership, inflates generated code, and makes it harder to
reason about which API surface each schema actually exposes.
### How
**1. Patch `@nestjs/graphql` to support schema-scoped type resolution**
- **(Already done)** Added a `resolverSchemaScope` option to
`GqlModuleOptions`, allowing each schema to declare a scope (e.g.
`'metadata'`)
- `ResolversExplorerService` now filters resolvers by a
`RESOLVER_SCHEMA_SCOPE` metadata key, so each schema only sees its own
resolvers
- `GraphQLSchemaFactory` now performs a **reachability walk**
(`computeReachableTypes`) starting from scoped resolver return types and
arguments, only including types that are transitively referenced —
handling unions, interfaces, and prototype chains
- Type definition storage and orphaned reference registry are cleared
between schema builds to prevent cross-contamination
**2. Register `ClientConfig` as orphaned type in metadata schema**
Since `ClientConfig` is needed in the metadata schema but not directly
returned by a resolver, it's explicitly declared via
`buildSchemaOptions.orphanedTypes`.
**3. Regenerate frontend types and fix imports**
- `generated/graphql.ts` shrank by ~2,700 lines (types moved to where
they belong)
- `generated-metadata/graphql.ts` gained types like `ClientConfig` that
were previously missing
- ~500 frontend files updated to import from the correct generated file
## Fix resolver schema leaking between `/metadata` and `/graphql`
endpoints
### Summary
- Patch `@nestjs/graphql` to support a `resolverSchemaScope` option that
filters resolvers at both schema generation and runtime, preventing
cross-endpoint leaking
- Introduce `@CoreResolver()` and `@MetadataResolver()` decorators to
explicitly scope each resolver to its endpoint
- Move most resolvers (auth, billing, workspace, user, etc.) to the
metadata schema where the frontend expects them; only workflow and
timeline calendar/messaging resolvers remain on `/graphql`
- Fix frontend `SSEQuerySubscribeEffect` to use the default (metadata)
Apollo client instead of the core client
### Problem
NestJS GraphQL's module-based resolver discovery traverses transitive
imports, causing resolvers from `/metadata` modules to leak into the
`/graphql` schema and vice versa. This made the schemas unpredictable
and tightly coupled to module import order.
### Approach
- Added `resolverSchemaScope` to `GqlModuleOptions` via a patch on
`@nestjs/graphql`, filtering in both `filterResolvers()` (runtime
binding) and `getAllCtors()` (schema generation)
- Each resolver is explicitly decorated with `@CoreResolver()` or
`@MetadataResolver()`
- Organized decorator, constant, and type files under `graphql-config/`
following project conventions
Core GQL Schema: (see: no more fields!)
<img width="827" height="894" alt="image"
src="https://github.com/user-attachments/assets/668f3f0f-485e-43f0-92be-4345aeccacb6"
/>
Metadata GQL Schema (see no more getTimelineCalendarEventsFromCompany)
<img width="827" height="894" alt="image"
src="https://github.com/user-attachments/assets/443913db-e5fe-4161-b0e7-4a971cc80a71"
/>
⚠️ **AI-generated PR — not ready for review** ⚠️
cc @FelixMalfait
---
## Changes
### System prompt improvements
- Explicit skill-before-tools workflow to prevent the model from calling
tools without loading the matching skill first
- Data efficiency guidance (default small limits, use filters)
- Pluralized `load_skill` → `load_skills` for consistency with
`load_tools`
### Token usage reduction
- Output serialization layer: strips null/undefined/empty values from
tool results
- Lowered default `find_*` limit from 100 → 10, max from 1000 → 100
### System object tool generation
- System objects (calendar events, messages, etc.) now generate AI tools
- Only workflow-related and favorite-related objects are excluded
### Context window display fix
- **Bug**: UI compared cumulative tokens (sum of all turns) against
single-request context window → showed 100% after a few turns
- **Fix**: Track `conversationSize` (last step's `inputTokens`) which
represents the actual conversation history size sent to the model
- New `conversationSize` column on thread entity with migration
### Workspace AI instructions
- Support for custom workspace-level AI instructions
---------
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
https://github.com/user-attachments/assets/c4e63edb-6e98-44bc-841f-ee110ae712d4
How it works
- `manifest.json` files are now committed when apps are published to our
repo
- to display available apps, from the server, we read into our github
repo, using a pod-scoped cache
- feature flagged
- app installation will be behind permission gate MARKETPLACE_APPS
Limitations and what is yet to develop
- content and settings tabs
- installed apps tab
- app installation
- test and potentially fix reading from .manifest.json once [Reshape
manifest
structure](https://github.com/twentyhq/core-team-issues/issues/2183) is
done. additional work is expected on assets notably. (couldnt properly
do it here as manifest.json will only be committed after this pr)
- we only read in community/ folder for now - we may want tochange that
- the cache is rather artisanal for now and scoped by pod - we may want
to change that
- Fix a race condition due to a page change effect re-trigger which
reset the focus stack (A user had a bug which happened when he opened
the command menu quickly after the page load. The focus stack was reset
and the hotkeys listening were the one from the table instead of the one
from the command menu)
- Fix the focus id in the side panel input which prevented using the
hotkeys
This PR is a follow-up of https://github.com/twentyhq/twenty/pull/16966
and implements a new mechanism to handle SSE events.
It creates only one event stream per browser tab, then use mutations to
tell the backend which query to listen to, without re-mounting the event
stream connexion.
Then each event that comes from this unique subscription is then
dispatched in a new JavaScript CustomEvent, per queryId, on which
specific hooks add an event listener.
This PR introduces the generic tooling as well as the handling of update
events on table.
## Context
The feature has not been maintained for more than a year and was never
officially launched.
This PR removes its code due to the upcoming refactoring of the
sync-metadata that will break its logic if not handled properly and it
would be too costly for the time being.
TODO:
- remove isRemote from object/field metadata
## Summary
This PR enables serverless functions to be exposed as AI tools, allowing
them to be used by AI agents.
### Changes
- Added new `SERVERLESS_FUNCTION` tool category
- Added `toolDescription`, `toolInputSchema`, and `toolOutputSchema`
fields to serverless functions
- Created database migration for the new schema columns
- Added tool index query and resolver for fetching available tools
- Added Settings AI page tabs (Skills, Tools, Settings) with new tools
table
- Added utility to convert tool schema to JSON schema format
- Updated frontend to display tools in the settings page
### Implementation Details
- Serverless functions can now define tool metadata (description,
input/output schemas)
- These functions are automatically registered in the tool registry
- The tool index endpoint allows querying available tools with their
schemas
- Settings page now has a dedicated Tools tab showing all available
tools
## Summary
This PR introduces a Skills system for AI agents, inspired by the [Agent
Skills specification](https://agentskills.io/specification).
## Changes
### Backend
- **SkillEntity**: New database entity with migration for storing skills
- **V2 Sync Mechanism**: Implemented FlatSkill, builders, validators,
and action handlers following the v2 flat entity pattern
- **Standard Skills**: Pre-defined skills (workflow-building,
data-manipulation, dashboard-building, metadata-building, research,
code-interpreter, xlsx, pdf, docx, pptx)
- **GraphQL API**: CRUD operations for skills with proper guards and
permissions
- **Workspace Cache**: Integrated skills into the workspace cache system
### Frontend
- **Skills Table**: Searchable table in AI settings showing all skills
- **Skill Form**: Create/edit page with Label (primary), Description,
and Content (markdown editor)
- **API Name**: Following existing patterns, name is derived from label
with advanced settings toggle for custom API names
- **Standard vs Custom**: Standard skills are read-only, custom skills
can be edited/deleted
## Key Design Decisions
- Skills are stored in the database (Salesforce-like approach) rather
than files
- Name is derived from Label by default (isLabelSyncedWithName pattern)
- Skills reference functions/files via @ mentions in markdown content
rather than explicit relations
- Standard skills are synced from code, custom skills are created via UI
## Screenshots
Skills table and form UI follow existing settings patterns.
## Testing
- [x] Lint passes
- [x] Typecheck passes
- [ ] CI tests
Félix Malfait