Adds intelligent routing system that automatically selects the best
agent for user queries based on conversation context.
### Changes:
- Added `routerModel` column to workspace table for configurable router
LLM selection
- Implemented `RouterService` with conversation history analysis and
agent matching logic
- Created router settings UI in AI Settings page with model dropdown
- Removed agent-specific thread associations - threads are now
agent-agnostic
- Added real-time routing status notification in chat UI with shimmer
effect
- Removed automatic default assistant agent creation
- Renamed GraphQL operations from agent-specific to generic (e.g.,
`agentChatThreads` → `chatThreads`)
---------
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
Co-authored-by: Félix Malfait <felix@twenty.com>
## What I've Implemented
I've added tooltips that show the original action type when you hover
over workflow step icons in the side panel. Now even if you rename an
action to something custom, you can still see what type of action it
actually is by hovering over the icon.
## The Problem This Solves
Before this change, once you renamed a workflow action (like changing
"Create Record" to "Add New Customer"), there was no way to tell what
the original action type was. This made it really confusing when
collaborating with others or when coming back to your own workflows
after some time - you couldn't tell if an action was a "Create Record"
or "Update Record" or something else.
## What Changed
Updated action type labels: Instead of showing just "Action" for all
record operations, the system now shows specific types like "Create
Record", "Update Record", "Delete Record", etc.
Added hover tooltips: When you hover over the action icons in the
workflow side panel, a tooltip appears showing the original action type,
even if you've renamed the step title.
## Before vs After
Before: All actions showed "Action" in the side panel, making it
impossible to distinguish between different types after renaming.
After: Each action shows its specific type in tooltips, so you always
know what you're working with.
This should make workflow management much clearer, especially when
multiple people are collaborating on the same workflow!
Resolves#14878
## Demo Video
See it in action:
https://www.loom.com/share/c0e0ec24e4524d0685b841e9ceb011d0?t=65&sid=dc05e58a-8869-4969-aa67-9772242fe697
---------
Co-authored-by: Thomas Trompette <thomas.trompette@sfr.fr>
Step 1 towards fixing issue #14976
Refactored WorkflowSendEmailBody to FormAdvancedTextFieldInput For its
reusability across application
- `useEmailEditor` is transformed to `useAdvancedTextEditor`: A
reuseable hook for managing the text editor state and functionality.
- `WorkflowSendEmailBody` is transformed to
`FormAdvancedTextFieldEditor`: A wrapper component for the advanced text
editor.
- `WorkflowSendEmailBody` is transformed to
`FormAdvancedTextFieldInput`: A component integrating the advanced text
editor into forms.
- `ImageBubbleMenu`, `LinkBubbleMenu`, and `TextBubbleMenu`: Contextual
menus for image, link, and text formatting options. are moved to
ui/components for access in FormAdvancedTextFieldInput
Additionally, the email action workflow component was updated to utilize
the new reuseable text editor, enhancing the user experience for
composing emails.
This update also includes storybook entries for the new components to
facilitate testing and documentation.
---------
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
Co-authored-by: Félix Malfait <felix@twenty.com>
# Implement "Tidy Up" Action for Workflow Diagram
**Task Link:** https://twill.ai/twentyhq/ENG/tasks/6
## Summary
This PR adds a "Tidy Up" action to the workflow diagram interface,
allowing users to automatically organize and clean up the layout of
workflow nodes and connections.
## Changes Made
- **Added new workflow action**: Created
`TidyUpWorkflowSingleRecordAction` component to provide a UI action for
tidying up workflow diagrams
- **Refactored tidy-up logic**: Extracted core tidy-up functionality
into a reusable `useTidyUp` hook, separating layout logic from workflow
version persistence
- **Updated action menu configuration**: Integrated the new tidy-up
action into `WorkflowActionsConfig` with proper permissions and keyboard
shortcuts
- **Enhanced right-click menu**: Updated
`WorkflowDiagramRightClickCommandMenu` to use the refactored tidy-up
hook
- **Improved hook architecture**: Simplified `useTidyUpWorkflowVersion`
to delegate layout calculations to the new `useTidyUp` hook
## Key Features
- Users can now trigger workflow diagram tidy-up from the action menu
- Consistent tidy-up behavior across both right-click menu and action
menu interfaces
- Better separation of concerns between layout calculation and data
persistence
<details>
<summary>📸 Screenshots</summary>
Playwright test screenshots captured during development:
### command-menu-with-tidy-up-workflow.png
### tidy-up-workflow-command-menu.png
### tidy-up-workflow-error-toast.png
</details>
---------
Co-authored-by: Twill <agent@twill.ai>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Thomas Trompette <thomas.trompette@sfr.fr>
This PR adds the components for advanced filters on chart settings.
It also refactors what's in common between this new command menu page
and the workflow advanced filters.
I did an acceptable design for empty node and iterators for release:
- use array field for iterator node. Added an util to stringify arrays
for backward compatibily
- remove icon for empty node
- allow to select a node on empty node selection
https://github.com/user-attachments/assets/b00037a8-aa1d-4784-b973-05973649b46e
When updating an active version, we :
- create a draft
- update the draft
Issue is that after the creation, workflow is shortly null. So the
component is re-rendered and the form data are lost.
Removing the check on the null allow to keep the form data.
Avoid fetching full steps and trigger for versions that are not the
current version. Because those won't be used anyway. Better for
performances.
Only difficulty was for the `createDraft` mutation. I needed to return
the full created version so I can store it in cache and use it as new
`currentVersion`. Otherwise the current version is considered as
incomplete for a short time, since workflow is fetched separately from
the current version.
---------
Co-authored-by: Charles Bochet <charles@twenty.com>
## Refactor: Prepare frontend record layouts for backend-driven
configuration
This PR simplifies and prepares the frontend record layout system for
eventual migration to backend-driven layouts, aligning the architecture
with the existing PageLayout system used for dashboards.
### Key Changes
**Architecture Improvements:**
- Created unified LayoutRenderingContext that works for both record
pages (with targetRecord) and dashboards (standalone)
- Replaced prop drilling with context-based data flow - cards access
targetRecord and isInRightDrawer via context hooks
- Introduced useTargetRecord() helper hook that provides type-safe
access to the current record
- Created generic CardRenderer component that handles configuration
injection and context guards uniformly
**Configuration System:**
- Converted all tab icons from React components to JSON-serializable
strings (e.g., Icon: IconCheckbox → icon: 'IconCheckbox')
- Added configuration field to cards, matching the widget configuration
pattern on the backend
- Created CardConfiguration types system similar to WidgetConfiguration
on the backend
- Moved widget-specific props (like showDuplicatesSection) into
configuration objects
**Code Organization:**
- Extracted visibility evaluation logic into reusable
evaluateTabVisibility() utility
- Organized layouts into dedicated files (one per object:
base-record-layout.ts, company-record-layout.ts, etc.)
- Renamed components to match their purpose: Notes → NotesCard,
Attachments → FilesCard, etc.
- Consolidated card rendering from registry object to direct
getCardComponent() function
**API Alignment:**
- Made ifNoReadPermissionObject an explicit part of TabVisibilityConfig
(follows if* naming convention)
- Removed redundant targetObjectNameSingular from tab-level (now derived
from visibility config)
- Card API now mirrors Widget API (both use type, configuration,
accessed via context)
## Problem
With Twenty usage growing, we are facing challenges on server side to
respond to the demand. The current bottleneck we are facing is server
CPU.
While investigating CPU performances, we figured out that loading all
relation fields was the biggest issue.
### About graphql query response size
**Example:** on `People Index` Table page:
we query: `person.company` and we query **all non relation fields** on
company relation field.
`{ person { company { id, name, domainName, employees, address ... } }`
However we only need **company.id, company.name and company.domainName**
to be able to render the Company Chip in the company column (RELATION)
in the table.
`{ person { company { id, name, domainName } }`
We initially assumed that the querying all non relation fields was not
an issue because it was not adding additional load on database (which is
also partially true: if a field is containing a huge json this will add
load on postgres as data needs to be transfered)
This assumption is wrong on CPU side:
- when loading one to many relations (company.people), this quickly ads
up and we end up with response of 20kB quite quickly, even without
adding any custom field on person.
- even when loading a many to one relation, if the user is storing big
data in a given field (note.body for instance, or workflowRun.state),
this starts also being an issue.
- Worst case scenario are starting to happen: on a production workspace
with 20 active workflows. Loading workflow table while displaying
workflow runs commands (20 workflows x 60 workflowRuns x
workflowRun.state which is a big JSON) result in a response of... 125MB.
### Why is it an issue?
When yoga (graphql engine) parses a response to send it back in server
response, it's using `JSON.parse `(or stringify depending on the case).
Parsing data is CPU instensive (as you need to validate, transform) and
this is more or less `O(n)`.
This means returning 125MB is very intense on the CPU and will likely
use 100% of the CPU for ~1sec in our production servers.
NodeJs (nodeV8) is single threaded. It's able to process multiple
requests in parallel but to do that, it will cut them in "microTasks"
and process each microTasks (that can belong to different requests) one
after the other. Exactly like a single threaded CPU would allocate some
time to a process and then to the next one, etc.
This means that this big response request will actually block the other
request.
As a result, all requests are slow and our infrastructure is multi
tenant so some customers are impacting others.
This is even a bigger issue when our health checks start to fail and
containers are starting being considered as unhealthy and killed by the
orchestrator
## How to fix this
1. On Front end side, we should only query what we need. In this PR, I'm
forcing the relations to only query: `id`, `labelIdentifier`,
`imageIdentifier`
2. (later) as we are API first, we also need to do something to protect
the servers from this side too. To do that, Graphql APIs can associate a
cost to each graphql field (maybe 1 for a test fixed, 2 for a JSON
field, 10 for relations, etc...) and we can throttle based on that.
## Changes in this PR
It's mainly about refactor the tooling to generate `RecordGqlFields`
(`generateDepthOneGqlFields`). this tooling is used to generate the list
of fields we want to query. Most of the time we want to query a record
with one level of nesting.
1) Refactor to remove duplicate code => `generateDepthOne` become
`generateDepth` and can handle both `depth = 0 | 1`
2) Introduce `generateDepthRecordGqlFieldsFromFields.ts` which is the
base and can give you a list of gqlFields based on FieldMetadataItems
3) Introduce `generateDepthRecordGqlFieldsFromObject` which is a
shortcut for an object
4) Introduce `generateDepthRecordGqlFieldsFromRecords` which is the
intersection between `generateDepthRecordGqlFieldsFromObject` and a
given record (useful for cache tooling)
5) Replace all usages + introduce a hook
`useGenerateDepthRecordGqlFieldsFromObject` to ease usage
This PR follows the multiSelect PR merged previously. It will enable
morph relation Many to One to be handled from the table, using a
singleSelect picker
Main point : I decided to change the singleSelect API to take an array
of **objectMetadataName** instead of only one to deal with both our
usecases.
---------
Co-authored-by: Charles Bochet <charles@twenty.com>
## Performance short term fix
Done in this PR:
workflowVersions and workflowRuns are heavy object. I'm preventing
loading their data when they are loaded as relations. This will reduce
the work on backend
## Long term fix
Todo:
We should make sure workflow data is ligher
**Issue**: https://github.com/twentyhq/core-team-issues/issues/1587
### Changes Made
- Modified scroll event handling to detect and limit two-finger scroll
gestures to the workflow canvas component
- Improved user experience by preventing accidental navigation or
zooming outside the intended canvas area
---------
Co-authored-by: Twill <agent@twill.ai>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
## Improvements
- Add logs to all gql operations and rest calls to help debug CPU issues
on the backend. These are temporary and should be removed
- Remove nested relations from workflowVersions load (used to add manual
triggers in the side bar). On some workspaces this call result in a
response of 4MB which is heavy on CPU
- Investigated Redis Usage ==> made a few improvements, we are should
still migrate to the new cache service once available
- investigated db calls in messaging / calendar fetch list + workflow
enqueue run cron jobs. Everything seems to be properly batched
This PR is the first part of the creation of the Chart editor.
https://github.com/user-attachments/assets/8b0af8ea-be41-4506-84cb-e40f5521cbf0
Done:
- Bar chart settings (except filters)
- Line chart settings (except filters)
In progress:
- Pie chart settings
- Number chart settings
- Gauge chart settings
Left to do:
- Loosen the backend validation to allow the user to save a partial
configuration, validate the graph configuration in the frontend and
display a error friendly message in the graph if the config is not
completed yet
- Implement the filter edition
- Finish the other graph types settings
## Context
Add routes to migration V2
- Resolvers
- Service v2
- Builder
- Validator
- Action runner
Next PR: Add to twenty-cli to sync routes with serverless
In this PR https://github.com/twentyhq/twenty/pull/14785 we deprecated
viewFilterOperand (from camelCase to capital snakeCase values), but we
had not migrated the existing workflow steps values. Hence they cannot
be executed!
Let's fix that by dealing with both operands, new and deprecated, to
mitigate the issue. Then we will add a command to migrate the values.
- add serverlessFunction schema in twenty-cli
- add trigger schema in twenty-cli
- update serverless function code save and get
- sync serverless function
In the BE we have stored filter operand as IS_EMPTY, IS_NOT_EMPTY, etc.
For some reason in the FE we were manipulating IsEmpty, IsNotEmpty, etc.
(maybe because they were used before in Views before they were moved to
core)
So we were converting the operands in the FE from IS to Is
(convertViewFilterOperandFromCore) to read and manipulate viewFilters,
and then back to BE version to send mutations etc., from Is to IS
(convertViewFilterOperandToCore).
The migration is now over, so we can remove and simplify that code.
(In the 1-5:migrate-views-to-core command we still do the migration from
Is format to IS format.)
Closes https://github.com/twentyhq/core-team-issues/issues/1560
If viewId is defined in a groupBy query, we want to apply all filters of
the view to the query.
This required to move a lot of code from twenty-front to twenty-shared
to convert the filters as stored in the db into graphql filters,
applying the right combinations between filters etc., which was
previously only done in the FE.
This PR does not handle any field filters, it will be done in a later pr
https://github.com/user-attachments/assets/d6c565eb-9a29-4830-9396-5f979c8caa7b
- Added a new component for manual trigger (mostly duplicated from
previous one). Will remove the old one once all data are migrated
- Updated schema output so the current item of the iterator can be typed
Todo left:
- migrate old triggers
- add an util to search iterator output. Today current item fields will
be displayed as not found
- set new manual triggers for workflow runs
- Added 'Cron will be triggered at UTC time' notice below trigger
interval dropdown
- Positioned correctly between dropdown and expression field to match
design
- Only shows when Custom CRON option is selected
---------
Co-authored-by: Félix Malfait <felix@twenty.com>
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
This PR should fix optimistic rendering issues on step updates:
- compute a diff for trigger and steps on mutations
- build an util that applies that diff (built a more robust version from
https://github.com/AsyncBanana/micropatch)
- apply diff in cache
