This PR adds a reload-safe onboarding Help Center generation status
path. Previously, generation progress was pushed through ActionCable,
which meant the onboarding UI could lose context after a page reload or
missed websocket event. The new endpoint exposes the current generation
id, raw Redis generation state, and Help Center article/category counts
so clients can recover state by fetching from the backend.
**What changed**
- Added an onboarding Help Center generation status endpoint.
- Persisted `help_center_generation_id` when generation is enqueued.
- Removed Help Center generation ActionCable broadcasts completely.
- Kept generation progress in Redis as the backend source of truth.
- Kept Help Center generation out of the onboarding hot path; the
existing onboarding controller does not start generation yet.
**How to test**
1. Start Help Center generation for an account.
2. Fetch the onboarding generation status endpoint and verify it returns
the generation id, Redis state, and article/category counts.
3. Reload the onboarding UI or client state and fetch again to confirm
progress can be recovered without relying on websocket events.
4. Verify skipped/completed generation states are reflected from Redis.
## Related PRs
- https://github.com/chatwoot/chatwoot/pull/14569
- https://github.com/chatwoot/chatwoot/pull/14568
- https://github.com/chatwoot/chatwoot/pull/14567
- https://github.com/chatwoot/chatwoot/pull/14619
- https://github.com/chatwoot/chatwoot/pull/14565 (Primary onboarding
PR)
## Description
Fixes WhatsApp embedded signup failing with "No WABA scope found in
token." Meta recently changed which scopes embedded-signup tokens carry
(whatsapp_business_manage_events instead of
whatsapp_business_management), which tripped a brittle scope-string
check. That check was redundant anyway — PhoneInfoService already
verifies the token's access to the specific WABA by calling its
/phone_numbers endpoint right before it. This removes the obsolete
TokenValidationService and relies on the functional check.
## Type of change
- [ ] Bug fix (non-breaking change which fixes an issue)
## Checklist:
- [ ] My code follows the style guidelines of this project
- [ ] I have performed a self-review of my code
- [ ] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [ ] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
Captain custom tools previously used their own hand-rolled `Net::HTTP`
request code. This moves them onto `SafeFetch`, the same shared
HTTP-fetching helper already used by webhooks, uploads, avatar imports,
and the Captain page crawler. Behavior for normal tools is unchanged —
they just now share one consistent path for host resolution, timeouts,
response size limits, and redirect handling.
**What changed**
- `HttpTool#execute_http_request` now delegates to `SafeFetch.fetch`
instead of building `Net::HTTP` requests by hand. Auth headers, basic
auth, metadata headers, JSON content-type, and the 1 MB response cap all
map onto `SafeFetch` options.
- Removed ~80 lines of bespoke request/validation plumbing from
`HttpTool`.
- The custom tools `test` endpoint now reads the response body string
directly (the executor returns the body rather than a response object).
**How to test**
1. Enable `custom_tools` (or `captain_integration_v2`) for an account.
2. Create a Captain custom tool pointing at a public HTTPS endpoint
(e.g. a test API).
3. Use the **Test** button in the tool form — you should get a success
result.
4. Run the tool from a Captain conversation and confirm the response is
returned/templated as before.
**Gotchas**
- **Local dev:** `SafeFetch` blocks requests to private/loopback
addresses by default. If you're testing a custom tool against a service
on `localhost` or a private IP during development, set
`SAFE_FETCH_ALLOW_PRIVATE_NETWORK=true` or the request will be rejected.
(This matches how the rest of `SafeFetch` already behaves locally.)
- **Test endpoint status field:** on success the `test` response now
reports `status: 200` rather than the exact 2xx code (201/204/etc.),
because `SafeFetch` signals success-vs-failure rather than exposing the
raw response. The UI only checks the 2xx range, so this is invisible
there — but worth knowing if anything consumes the API directly.
- **Non-2xx responses** still surface as an error (same as before), now
via `SafeFetch::HttpError`.
Updates the locked OAuth dependencies to patched versions so bundle
audit no longer reports the current OAuth advisories.
What changed
- Updated `oauth` from `1.1.0` to `1.1.6` for `GHSA-prq8-7wvh-44qh`.
- Updated `oauth2` from `2.0.9` to `2.0.22` for `GHSA-pp92-crg2-gfv9`.
- Accepted Bundler's required transitive lockfile updates for the
patched OAuth gems.
How to test
1. Run `bundle exec bundle audit update && bundle exec bundle audit
check -v` and confirm no vulnerabilities are reported.
2. Smoke test OAuth-based authentication and email integration flows.
Co-authored-by: Shivam Mishra <scm.mymail@gmail.com>
Recording and sending an audio message from a **Twilio WhatsApp** inbox
failed silently — Twilio rejected the media with delivery error `63019`
("Media failed to download") and the voice note never reached the
customer. This restores audio sending for Twilio WhatsApp (and
360dialog) inboxes.
Closes Regression from #14606
## How to reproduce
1. Open a conversation in a **Twilio WhatsApp** inbox.
2. Record a voice message in the reply box and send it.
3. Before this fix: the message fails to deliver and a
`Webhooks::TwilioDeliveryStatusJob` is enqueued with `ErrorCode: 63019`,
`ErrorMessage: "Media failed to download"`.
4. After this fix: the audio is recorded as MP3 and delivers normally.
## What changed
PR #14606 added WhatsApp **Cloud** voice notes, which require OGG/Opus.
It changed `audioRecordFormat` in `ReplyBox.vue` to return OGG for
`isAWhatsAppChannel` — but that getter is also `true` for Twilio
WhatsApp inboxes. The OGG handling (content-type normalization + the
`voice: true` flag) lives only in `WhatsappCloudService`, so Twilio
could not download/process the remuxed OGG file.
This change scopes OGG to `isAWhatsAppCloudChannel`. Twilio WhatsApp,
360dialog, and Telegram fall back to MP3 exactly as they did before the
PR.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-authored-by: Muhsin <12408980+muhsin-k@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Restrict the WhatsApp reauthorization flag to channels whose
provider_config source is 'embedded_signup'. Previously the view exposed
resource.channel.reauthorization_required? for all WhatsApp channels;
now it only returns true when (provider_config || {}).to_h['source'] ==
'embedded_signup' && resource.channel.reauthorization_required?. This
prevents showing reauthorization prompts for manual/API-key flows
(non-OAuth) and safely handles nil provider_config.
# Pull Request Template
## Description
Please include a summary of the change and issue(s) fixed. Also, mention
relevant motivation, context, and any dependencies that this change
requires.
Fixes #13553
## Type of change
Please delete options that are not relevant.
- [x] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality not to work as expected)
- [ ] This change requires a documentation update
## How Has This Been Tested?
Please describe the tests that you ran to verify your changes. Provide
instructions so we can reproduce. Please also list any relevant details
for your test configuration.
## Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] I have commented on my code, particularly in hard-to-understand
areas
- [x] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
---------
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
Co-authored-by: Muhsin <12408980+muhsin-k@users.noreply.github.com>
Conversation and inbox webhook payloads now include the account object
(`{ id, name }`) so integrations can reliably identify the account
without depending on nested message data.
## Closes
Closes#11753Closes#12442
## Why
Message, contact, and contact inbox webhook payloads already expose
`account`. Conversation and inbox webhook payloads were outliers, which
forced webhook consumers to infer account context from nested messages
or other fields that may not always be present.
## What changed
- Adds `account: { id, name }` to conversation webhook payloads.
- Adds webhook-specific inbox payload data with `account: { id, name }`
for inbox created/updated events.
- Keeps non-webhook conversation and inbox push payload behavior
unchanged.
## How to test
- Configure an account webhook for `conversation_created`, trigger a new
conversation, and confirm the webhook payload includes `account.id` and
`account.name`.
- Configure an account webhook for `inbox_created`, create a new inbox,
and confirm the webhook payload includes `account.id` and
`account.name`.
- Configure a webhook agent bot and update a conversation status, then
confirm the bot webhook payload includes the same account object.
---------
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
Co-authored-by: Sojan Jose <sojan@pepalo.com>
Inbound emails containing null bytes could fail while Chatwoot persisted
the incoming message, causing IMAP sync to drop and repeatedly retry the
same malformed email. This strips null bytes at the inbound mailbox
message persistence boundary so the rest of the email can be saved
normally.
Fixes
https://linear.app/chatwoot/issue/CW-7165/argumenterror-string-contains-null-byte-argumenterror
Sentry: https://chatwoot-p3.sentry.io/issues/7405946944/
## Test
- [x] added specs
- [x] manually tested the logic on prod for the inbox(email) which was
throwing the error
## What changed
- Sanitizes null bytes only while building attributes for inbound
mailbox message persistence.
- Uses the sanitized source ID for both duplicate lookup and message
creation.
- Adds regression coverage for the mailbox message creation path.
---------
Co-authored-by: Sony Mathew <2040199+sony-mathew@users.noreply.github.com>
Co-authored-by: Sony Mathew <sony@chatwoot.com>
Outbound Facebook Messenger replies were always sent with
`messaging_type: MESSAGE_TAG` and a `tag` of either `HUMAN_AGENT` or
`ACCOUNT_UPDATE` (fallback). `ACCOUNT_UPDATE` is reserved by Meta policy
for non-recurring account notifications (password resets, suspicious
activity, account-status changes), so general agent replies were getting
rejected by Facebook with "Invalid parameter" and risked page-level
penalties. After this fix, Facebook treats default replies as standard
`RESPONSE` messages within the 24-hour window, and only attaches
`HUMAN_AGENT` tagging when the operator opts in via
`ENABLE_MESSENGER_CHANNEL_HUMAN_AGENT`.
## Closes
<!-- Add the relevant issue link, e.g. Closes#1234 -->
## How to reproduce
1. Connect a Facebook page inbox.
2. Leave `ENABLE_MESSENGER_CHANNEL_HUMAN_AGENT` disabled (default).
3. From the dashboard, reply to an active Messenger conversation with a
regular/promotional message.
4. Before this PR: the request to Graph API includes
`messaging_type=MESSAGE_TAG&tag=ACCOUNT_UPDATE`. Facebook returns
"Invalid parameter" for content that does not match the `ACCOUNT_UPDATE`
use case, and the message is marked failed in Chatwoot.
5. After this PR: the request omits `messaging_type` and `tag`, so
Facebook treats it as a standard `RESPONSE` and accepts it within the
24-hour customer service window.
## What changed
- `app/services/facebook/send_on_facebook_service.rb` now mirrors the
Instagram service pattern: text and attachment params are built without
`messaging_type`/`tag` by default, and a shared `merge_human_agent_tag`
helper attaches `MESSAGE_TAG`/`HUMAN_AGENT` only when
`ENABLE_MESSENGER_CHANNEL_HUMAN_AGENT` is enabled.
- Removed the unused `sent_first_outgoing_message_after_24_hours?`
helper (dead code with no callers).
- Updated `spec/services/facebook/send_on_facebook_service_spec.rb`:
dropped `ACCOUNT_UPDATE` expectations, added a spec asserting no
`messaging_type`/`tag` is sent by default, and tightened the HUMAN_AGENT
spec to verify both `messaging_type` and `tag`.
## Operator note
Operators who were relying on the prior `ACCOUNT_UPDATE` fallback to
bypass the 24-hour window were already in violation of Meta policy and
likely seeing send failures. They should enable
`ENABLE_MESSENGER_CHANNEL_HUMAN_AGENT` (and request the Human Agent
permission from Meta) for the 7-day extended window
---------
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
Co-authored-by: Muhsin <12408980+muhsin-k@users.noreply.github.com>
# Pull Request Template
## Description
Adds a Contact condition to the conversation advanced filter so agents
can search for an existing contact and filter conversations by
`conversations.contact_id`.
Fixes
[CW-7239](https://linear.app/chatwoot/issue/CW-7239/contact-filter-for-conversations)
## Type of change
- [ ] Bug fix (non-breaking change which fixes an issue)
- [x] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality not to work as expected)
- [ ] This change requires a documentation update
## How Has This Been Tested?
- `eval "$(rbenv init -)" && bundle exec rspec
spec/services/conversations/filter_service_spec.rb`
- `pnpm exec vitest --no-watch --no-cache --no-coverage --logHeapUsage
app/javascript/dashboard/components-next/filter/helper/filterHelper.spec.js
app/javascript/dashboard/store/modules/conversations/helpers/specs/filterHelpers.spec.js
app/javascript/dashboard/helper/specs/customViewsHelper.spec.js
app/javascript/dashboard/helper/specs/filterQueryGenerator.spec.js`
- `pnpm eslint`
- `eval "$(rbenv init -)" && bundle exec rubocop
spec/services/conversations/filter_service_spec.rb`
## Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
---------
Co-authored-by: iamsivin <iamsivin@gmail.com>
Co-authored-by: Sivin Varghese <64252451+iamsivin@users.noreply.github.com>
Updates Puma to the patched 7.2 line so bundle-audit no longer reports
the high-severity Puma advisories affecting the current lockfile.
Fixes: CVE-2026-47736, CVE-2026-47737
Closes: N/A
## Why
The refreshed ruby-advisory-db flagged Puma 6.4.3 for two high-severity
PROXY Protocol v1 parser advisories. The advisory-recommended fixed
versions are the Puma 7.2 patch line or Puma 8, so this keeps the update
to the smaller fixed line.
## What changed
- Constrains Puma to `~> 7.2`, `>= 7.2.1`
- Updates the lockfile from Puma 6.4.3 to 7.2.1
- Updates nio4r from 2.7.3 to 2.7.5 as the transitive Puma dependency
## Validation
- `bundle exec bundle-audit check` -> `No vulnerabilities found`
- `bundle exec rails runner 'puts "Rails #{Rails.version}; Puma
#{Puma::Const::PUMA_VERSION}; Rack #{Rack.release}"'` -> `Rails 7.1.5.2;
Puma 7.2.1; Rack 3.2.6`
- `bundle exec rails server -p 3051` booted Puma 7.2.1 successfully
- Smoke checked `/`, `/api`, unknown-route 404, and `/cable` WebSocket
upgrade
# Pull Request Template
## Description
Fixes:
https://linear.app/chatwoot/issue/CW-7167/label-suggestions-bad-ux
## Type of change
Please delete options that are not relevant.
- [x] Bug fix (non-breaking change which fixes an issue)
## How Has This Been Tested?
Please describe the tests that you ran to verify your changes. Provide
instructions so we can reproduce. Please also list any relevant details
for your test configuration.
## Checklist:
- [ ] My code follows the style guidelines of this project
- [ ] I have performed a self-review of my code
- [ ] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [ ] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
Extracts the Meta JS SDK logic for the WhatsApp embedded-signup and
Facebook Page connect flows out of their settings components into two
reusable composables. No behavior change — the settings inbox-creation
flows work exactly as before; this just makes the SDK orchestration
reusable (an upcoming onboarding PR consumes them) and adds unit
coverage.
## What changed
- `useWhatsappEmbeddedSignup` — owns the embedded-signup popup (SDK
load, FB.login, and the auth-code / postMessage race).
`WhatsappEmbeddedSignup.vue` now consumes it.
- `useFacebookPageConnect` — owns FB.login (page scopes) + page fetch,
with SDK preload split from login so the popup opens within the click's
activation window. `Facebook.vue` now consumes it.
- Adds unit specs for both composables.
## Related PRs
- https://github.com/chatwoot/chatwoot/pull/14569
- https://github.com/chatwoot/chatwoot/pull/14568
- https://github.com/chatwoot/chatwoot/pull/14567
- https://github.com/chatwoot/chatwoot/pull/14649
- https://github.com/chatwoot/chatwoot/pull/14565 (Primary onboarding
PR)
Instagram webhooks use a Redis mutex to protect the first-message path
for a contact/account pair. When multiple webhook events for the same
Instagram contact arrive at nearly the same time, they can all enter the
“find or create” flow together.
There are two pieces involved:
`ContactInbox` maps the external Instagram scoped user id to a Chatwoot
contact inside an inbox. That side already has a unique `(inbox_id,
source_id)` index and retry handling, so duplicate contact-inbox
creation is mostly protected at the database layer.
`Conversation` creation is more fragile. The message builder looks for
an existing active conversation for the contact/inbox and creates one
when none is found. If two first-message jobs run concurrently, both can
see “no active conversation yet” and both can create a conversation. The
mutex exists to bump those jobs apart long enough for the first one to
create the conversation, so the next one appends to it instead of
creating a duplicate.
In production, the old behavior could turn that race dampener into a
drop path. Once `Webhooks::InstagramEventsJob` exhausted its lock
retries, ActiveJob stopped retrying with
`MutexApplicationJob::LockAcquisitionError`. Since ActiveJob retries are
not FIFO, later jobs could still win the lock while older jobs kept
getting deferred until exhaustion.
## Mutex Application Job Changes
This adds `retry_on_lock_conflict` as a small wrapper around the
existing `retry_on LockAcquisitionError` pattern.
The new API keeps the default behavior intact, but lets jobs explicitly
define what should happen after lock retry exhaustion:
```ruby
retry_on_lock_conflict wait: 1.second,
attempts: 3,
on_exhaustion: :process_without_lock
```
The fallback receives the original job arguments, so job classes do not
need to reach into ActiveJob internals like `arguments.first`.
## Instagram Fallback
Instagram now processes the webhook payload even after the mutex retry
window is exhausted.
This matches what the mutex was meant to do in the first place: bump
concurrent events apart long enough to avoid duplicate conversation
creation, not permanently block or drop webhook events. If a job cannot
acquire the lock after a few retries, it continues through the normal
Instagram processing path without the mutex.
## Retry And Lock Tuning
The Instagram lock retry window now uses deterministic backoff:
```ruby
retry_on_lock_conflict wait: ->(executions) { executions.seconds },
attempts: 3,
on_exhaustion: :process_without_lock
```
The lock TTL is also set explicitly:
```ruby
with_lock(key, 3.seconds)
```
This matters because Rails treats `attempts` as total executions, not
retries after the first execution. With a fixed `wait: 1.second` and
`attempts: 3`, the exhaustion handler can run roughly two seconds after
the first lock conflict. That is earlier than a 3-second lock TTL, so
the fallback could process without the lock while the original mutex is
still valid. That would reopen the duplicate-conversation race the lock
is meant to dampen.
Using a proc wait makes the timing predictable and avoids Rails jitter
for this retry path. The first conflict waits about 1 second, the second
waits about 2 seconds, and the final attempt happens around the 3-second
mark. That lines the retry window up with the Redis lock TTL before
falling back to `process_without_lock`.
The protected race window is intentionally small. `ContactInbox`
creation already has database uniqueness protection, while conversation
creation is the softer `find active conversation || create` path. We
only need to give the first job enough time to create the conversation
state that later jobs should reuse.
The mutex still does not preserve message order, and ActiveJob retries
are not FIFO. A longer retry window would mostly add latency for hot
Instagram contacts without making ordering more correct. After the lock
TTL has elapsed, processing without the lock is the better tradeoff than
dead-lettering customer messages.
WhatsApp template variables like `{{contact.first_name}}` previously
resolved to an empty string whenever a contact or agent had a
single-word name (e.g. "Petterson"). Because the variable came back
blank, WhatsApp template sends would either fail validation or get stuck
indefinitely in the "Sending" state. With this fix, `first_name` falls
back to the full single-word name, so templates render correctly for
everyone regardless of how many words their name has.
## Closes
- Fixes#13222
## How to reproduce
1. Create a contact with a single-word name (e.g. `Petterson`).
2. Configure a WhatsApp template action that uses
`{{contact.first_name}}`.
3. Send the template message.
4. **Before:** the variable resolves to an empty string and the message
fails / stays in "Sending".
**After:** the variable resolves to `Petterson` and the message sends.
## What changed
- Removed the "name must have 2+ words" guard from `first_name` in
`ContactDrop` and `UserDrop`, so a single-word name is returned (still
capitalized).
- Capitalization behavior introduced in #6758 is preserved; only the
single-word edge case changes.
- `last_name` still returns `nil` for single-word names, which is the
expected behavior.
- Added specs covering single-word names for both `ContactDrop` and
`UserDrop`.
---------
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
Co-authored-by: Muhsin <12408980+muhsin-k@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
Portals can now override their name, page title and header text per
locale, with a fallback chain of locale override → default locale → base
value. Overrides are stored under portal.config.locale_translations and
validated with a JSON schema.
Editing is exposed as a "Localize content" action in each non-default
locale's menu on the Locale page, and all public-facing portal views
(classic + documentation layouts) render the localized values. The live
chat widget is also loaded in the active portal locale.
<img width="494" height="395" alt="Screenshot 2026-06-03 at 2 48 59 PM"
src="https://github.com/user-attachments/assets/e55a06e8-f20f-4d8a-9352-92fde28a9a19"
/>
https://github.com/user-attachments/assets/f5affbd2-7ad4-415a-b376-57c759fc2aaa
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
## Description
Migrates Firecrawl from the v1 to the v2 API.
`Captain::Tools::FirecrawlService` now targets `api.firecrawl.dev/v2`,
with the request body updated to match the v2 schema.
> Disclosure: I work at Firecrawl.
Fixes # (n/a)
## Type of change
- [x] Bug fix (non-breaking change which fixes an issue)
## How Has This Been Tested?
Updated
`spec/enterprise/services/captain/tools/firecrawl_service_spec.rb` to
assert the v2 endpoint and request body.
## Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
---------
Co-authored-by: Aakash Bakhle <48802744+aakashb95@users.noreply.github.com>
# Pull Request Template
## Description
We need to pass on trace level attributes down to the spans inside them
like tool calls, observations, etc.
This way, we can filter observations based on trace level attributes.
## Type of change
- [x] Bug fix (non-breaking change which fixes an issue)
## How Has This Been Tested?
Please describe the tests that you ran to verify your changes. Provide
instructions so we can reproduce. Please also list any relevant details
for your test configuration.
Attributes added to observation metadata for easy filtering
<img width="1327" height="708" alt="image"
src="https://github.com/user-attachments/assets/8f1d1bf8-cde4-481d-a2c2-7920ad2fc52e"
/>
added a `generation_stage` to differentiate llm_calls that call tools vs
those that generate a `final_response`
<img width="1806" height="968" alt="CleanShot 2026-06-03 at 15 11 09@2x"
src="https://github.com/user-attachments/assets/db1fa8e0-7f2d-404b-a719-27a16d400442"
/>
propagated attributes to tool calls for future use
<img width="903" height="517" alt="image"
src="https://github.com/user-attachments/assets/edc61ce8-93db-465c-a66e-043138e2dc15"
/>
## Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
- [x] Any dependent changes have been merged and published in downstream
modules
## Linear Ticket
https://linear.app/chatwoot/issue/CW-7250/add-voice-calls-to-self-hosted-super-admin-feature-list
## Description
Adds a **Voice Calls** card to the Super Admin → Settings → Features
showcase so self-hostedadmins can see at a glance whether voice calling
is available on the installation.
## Type of change
- [ ] New feature (non-breaking change which adds functionality)
## Screenshots?
<img width="1512" height="843" alt="Screenshot 2026-06-03 at 3 15 22 PM"
src="https://github.com/user-attachments/assets/988db14b-fef1-4b72-a55e-c7a8884521dd"
/>
## Checklist:
- [ ] My code follows the style guidelines of this project
- [ ] I have performed a self-review of my code
- [ ] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [ ] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
## Linear Ticket
https://linear.app/chatwoot/issue/CW-7260/update-the-voice-call-ringtone
## Description
Updates the incoming voice call ringtone. The dashboard call widget now
plays the new tone while an inbound call is unanswered, replacing the
previous bell sound.
## Type of change
- [x] New feature (non-breaking change which adds functionality)
## Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] I have commented on my code, particularly in hard-to-understand
areas
- [x] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
- [x] Any dependent changes have been merged and published in downstream
modules
## Linear Ticket
https://linear.app/chatwoot/issue/CW-7264/add-mute-button-in-twilio-calls
## Description
Adds mute support for Twilio voice calls. Previously the mute button was
shown only for WhatsApp calls (which toggle the local mic track in the
browser), so Twilio calls had no way to mute. The call widget now routes
mute by provider: WhatsApp continues to toggle the local mic track,
while Twilio uses the Voice SDK connection's native `mute()`. The mute
button is now shown for any active call, and unmute works symmetrically.
## Type of change
- [x] New feature (non-breaking change which adds functionality)
## Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] I have commented on my code, particularly in hard-to-understand
areas
- [x] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
## Linear Ticket
https://linear.app/chatwoot/issue/CW-7261/call-window-theme-alignment
## Description
Updates the call window's reject/end button to use the hang-up handset
icon (the same handset as the accept button, rotated) instead of the
phone-with-X, matching the design.
## Type of change
- [x] New feature (non-breaking change which adds functionality)
## Screenshots
<img width="346" height="135" alt="Screenshot 2026-06-03 at 3 41 10 PM"
src="https://github.com/user-attachments/assets/4d507141-332e-436a-a6ec-516964a31013"
/>
## Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] My changes generate no new warnings
Fetch IMAP message content using `BODY.PEEK[]` instead of `RFC822` to
avoid provider-specific parser failures while preserving unread state.
This also applies the existing SMTP timeout configuration to custom SMTP
email-channel replies, so provider SMTP responses have enough time to
complete.
Fixes: https://github.com/chatwoot/chatwoot/issues/12762
## Why
Some IMAP providers can return responses for `FETCH RFC822` that Ruby
`net-imap` fails to parse with:
`Net::IMAP::ResponseParseError: unexpected RPAR (expected ATOM or NIL)`
We reproduced this with iCloud IMAP. Authentication, `INBOX` selection,
and header fetches worked, but fetching full message content with
`RFC822` failed before Chatwoot received a `Mail::Message`.
The same mailbox successfully returned full message content when fetched
with `BODY.PEEK[]`.
> During end-to-end iCloud validation, inbound fetch worked after the
IMAP change, but outbound replies through the custom SMTP settings could
still fail with a socket read timeout. The OAuth SMTP path already used
explicit SMTP timeout values; the custom SMTP path was relying on mailer
defaults instead.
## What this change does
- Replaces the full message fetch from `RFC822` to `BODY.PEEK[]`
- Reads the returned message content from `BODY[]`, which is how
`net-imap` exposes the response attribute
- Keeps the existing `BODY.PEEK[HEADER]` header-fetch behavior unchanged
- Applies `SMTP_OPEN_TIMEOUT` and `SMTP_READ_TIMEOUT` to custom SMTP
email-channel replies
- Defaults custom SMTP reply delivery to `open_timeout: 15` and
`read_timeout: 30`
- Updates IMAP service specs for standard and Microsoft IMAP fetch flows
- Updates mailer specs for custom SMTP timeout settings
`BODY.PEEK[]` is preferable here because it fetches the full message
content without marking messages as read.
## Validation
- Configured a local email inbox against iCloud IMAP and SMTP
- Confirmed `FETCH RFC822` reproduces `Net::IMAP::ResponseParseError:
unexpected RPAR (expected ATOM or NIL)`
- Confirmed `BODY[]` and `BODY.PEEK[]` fetch the same mailbox
successfully
- Confirmed Chatwoot imports iCloud messages after the IMAP change
- Sent two outbound replies from the Chatwoot UI through iCloud SMTP
after applying the timeout settings
- Confirmed both UI-created outbound messages were marked `sent`, had
iCloud SMTP `source_id` values, and had no `external_error`
- Ran `bundle exec rspec spec/services/imap/fetch_email_service_spec.rb
spec/services/imap/microsoft_fetch_email_service_spec.rb`
- Ran `bundle exec rspec spec/mailers/conversation_reply_mailer_spec.rb`
# Pull Request Template
## Description
This PR adds support for inline image uploads in the reply editor for
Email and Website (chat widget) channels.
Agents can now insert images inline between text and resize them
directly in the editor by dragging the bottom corner, similar to the
help center editor experience.
Image sizes are preserved through markdown using the `cw_image_width`
URL param and render correctly in both outgoing emails and chat widget
messages.
Agents can also paste copied images directly into Email or Website
replies using **Shift+Cmd+V** (Shift+Ctrl+V on Windows/Linux). The image
gets inserted inline at the cursor position and supports resizing just
like uploaded images. Regular **Cmd+V / Ctrl+V** behavior remains
unchanged and continues to add images as attachments, so both inline and
attachment flows are supported.
### Prosemirror repo PR:
https://github.com/chatwoot/prosemirror-schema/pull/48
Fixes
https://linear.app/chatwoot/issue/CW-7133/inline-images-in-live-chat-and-emailhttps://linear.app/chatwoot/issue/CW-7225/ghsa-8j9w-jppp-xcfc-html-attribute-injection-via-unvalidated-cw-image
## Type of change
- [x] New feature (non-breaking change which adds functionality)
## How Has This Been Tested?
### Screencast
https://github.com/user-attachments/assets/a928f852-ab15-413a-9d35-6ea69b718ecf
<img width="414" height="654" alt="image"
src="https://github.com/user-attachments/assets/205e0729-8f2d-4cc5-9c55-7696f032eca4"
/>
## Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
---------
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
## Summary
`Whatsapp::IncomingMessageBaseService#attach_location` builds a
`fallback_title` by concatenating `location['name']` and
`location['address']` with no length cap, then stores it directly into
`Attachment#fallback_title`. `ApplicationRecord` enforces a generic
255-character limit on string columns, so any WhatsApp location whose
`"#{name}, #{address}"` exceeds 255 chars (a common case for Google
Places that include a long full address) raises
`ActiveRecord::RecordInvalid` deep inside the Sidekiq job. The message
and attachment INSERTs are part of the same transaction, so the whole
thing rolls back. Sidekiq retries once; the retry dedup-skips the wamid
silently and exits without an error. **Result: the message is
irrecoverably lost — no row in `messages`, no entry in the UI, no
outgoing webhook, no clue for the operator.**
Confirmed in `v4.13.0`, `v4.14.0`, and `develop` (commit `f33e469`,
2026-05-20). No upstream issue found before opening this PR.
## How to reproduce
1. From WhatsApp, share a Google Place whose `name + ", " + address` is
> 255 chars. The Spanish business address `Gremi de Fusters, 33,
Edificio VIP Asima, Piso 2, Local 2, Norte, 07009 Polígon industrial de
Son Castelló, Illes Balears, España` (132 chars) used as both `name` and
`address` is enough.
2. Sidekiq logs:
```
ERROR ActiveRecord::RecordInvalid: Validation failed:
Attachments fallback title is too long (maximum is 255 characters)
```
3. The `messages` table has no row. The conversation UI shows nothing
for that timestamp.
4. The first retry "Performed" successfully but creates nothing — the
dedup-by-source-id silently swallows the failure.
## Fix
Cap the existing concatenated title at 255 chars via `.first(255)`.
Minimal change, no behavioural difference for any message shorter than
the limit, prevents the silent data loss for any longer ones.
```diff
- location_name = location['name'] ? "#{location['name']}, #{location['address']}" : ''
+ location_name = (location['name'] ? "#{location['name']}, #{location['address']}" : '').first(255)
```
## Alternatives considered
- **Increase the validation limit on `Attachment#fallback_title`**: more
invasive; would touch other inbound channels and possibly require a DB
column change.
- **Use `name` alone (no concat)**: cleaner semantically (in many real
payloads `name == address`), but changes user-visible behaviour. Left as
a follow-up if desired.
- **Truncate with ellipsis**: cosmetic only; deferred.
This PR is intentionally minimal so it can be merged on its own.
---------
Co-authored-by: Sony Mathew <sony@chatwoot.com>
Co-authored-by: Sony Mathew <2040199+sony-mathew@users.noreply.github.com>
Fixes#9775
## Description
This fixes a repeated admin consent loop in the Microsoft OAuth flow
when connecting a Microsoft email inbox.
Chatwoot was always sending `prompt=consent` in the Microsoft
authorization URL. In the current code path, this parameter is only used
when building the authorization URL and is not required by the callback,
token exchange, token persistence, or refresh flow.
By removing the forced consent prompt, the OAuth flow can proceed
normally without repeatedly sending users back through the admin consent
screen.
## What changed
- removed `prompt: 'consent'` from the Microsoft authorization URL
- added a regression assertion to ensure `prompt` is not included in the
generated URL
## Why this is safe
- `redirect_uri`, `scope`, and `state` remain unchanged
- callback and token exchange flow remain unchanged
- refresh token flow remains unchanged
- no other part of the current Microsoft inbox flow depends on forcing a
consent screen
## Testing
- updated controller spec to assert that the generated authorization URL
does not include `prompt`
The TipTap/ProseMirror editor stores agent messages with trailing
paragraph nodes that produce trailing newlines (e.g. \`\n\n\n\`) in the
\`content\` field. While Chatwoot's native channel delivery already
handles this, webhook payloads and API responses were returning raw
content with trailing whitespace — causing visible blank space below
messages in every external integration that consumes Chatwoot webhooks
(WhatsApp via Evolution API, Telegram bots, custom webhook consumers).
Closes#13459
## Root cause
\`Messages::WebhookContentNormalizer\` already strips CommonMark hard
line breaks (\`\\\` + newline) for webhook consumers, but it did not
strip trailing whitespace. All webhook and API responses flow through
this normaliser, so it is the single correct place to apply the fix
without touching stored data.
## What changed
Added \`.rstrip\` to \`Messages::WebhookContentNormalizer.normalize\`:
\`\`\`ruby
# before
text.gsub(/\\\r?\n/, "\n")
# after
text.gsub(/\\\r?\n/, "\n").rstrip
\`\`\`
## Trade-offs considered
| Option | Decision |
|---|---|
| \`before_save\` on \`Message\` model | Would clean stored data but is
a broader change affecting all message creation paths and would require
a data migration for existing records. Out of scope for this bug. |
| Trim in each channel's send path | DRY violation — many channels, each
would need the same patch. |
| Fix at normaliser level (chosen) | Single location, only affects
webhook/API output, zero risk to stored data or native channel delivery.
|
**Known limitation:** existing messages in the database still have
trailing newlines in storage. They will be delivered correctly through
webhooks after this fix, but a follow-up migration could clean stored
content if needed.
## How to reproduce
1. Send an agent reply from the Chatwoot UI
2. Inspect the \`content\` field of the outgoing \`message_created\`
webhook payload
3. Observe trailing \`\n\n\n\` after the message text
After this fix, the \`content\` field is trimmed before delivery.
---------
Co-authored-by: Ramalau Debeila <rdebeila@datacentrix.co.za>
Co-authored-by: Sivin Varghese <64252451+iamsivin@users.noreply.github.com>
Updates dashboard, widget, and backend locale files with the latest
translation sync from Crowdin.
## Closes
N/A
## What changed
- Refreshes translated dashboard JSON locale files across supported
languages.
- Adds the latest backend Help Center/public portal locale strings.
- Keeps the branch current with `develop` and resolves the `ar`, `fr`,
and `pt_BR` locale conflicts.
## Validation
- Parsed all changed JSON and YAML locale files successfully.
- Checked for leftover merge conflict markers.
- Ran `git diff --check`.
---------
Co-authored-by: Sojan Jose <sojan@pepalo.com>
This reverts #14610 so conversation unread counts are again controlled
by the `conversation_unread_counts` feature flag across the API,
ActionCable broadcasts, notifier/listener paths, and dashboard sidebar
fetching.
## Closes
- None
## What changed
- Restores feature-flag checks for conversation unread count reads and
broadcasts.
- Restores the dashboard feature flag constant and sidebar/store
behavior for disabled unread counts.
- Restores the specs that cover disabled-feature behavior.
## How to test
- In an account with `conversation_unread_counts` enabled, verify
sidebar unread counts are fetched and updated in real time.
- Disable `conversation_unread_counts` for the account and verify unread
count requests/broadcasts are skipped.
Translate missing Simplified Chinese live-chat widget availability and
reply-time strings so visitors using `zh_CN` no longer see English
fallbacks in offline and pre-chat states.
## Closes
N/A
## How to test
- Open the widget with `Widget Locale = 中文 (zh_CN)` and set the team
offline. The availability card should stay in Chinese.
- Configure working hours so the widget shows minute, hour, tomorrow,
and day-specific return states. The `{time}`, `{n}`, and `{day}`
placeholders should render correctly.
## What changed
- Translated missing `THUMBNAIL.AUTHOR.NOT_AVAILABLE`,
`TEAM_AVAILABILITY.BACK_AS_SOON_AS_POSSIBLE`, and `REPLY_TIME.*` strings
in `app/javascript/widget/i18n/locale/zh_CN.json`.
- Kept the scope limited to Simplified Chinese.
Co-authored-by: Sojan Jose <sojan@pepalo.com>
> Reopened from #13613, now from a personal fork
(`gabrieljablonski/chatwoot`) so maintainers can push edits —
organization-owned forks don't support "Allow edits from maintainers".
The previous PR is closed in favor of this one; same commits, same diff.
## Description
This PR adds support for sending voice messages (voice notes) through
the WhatsApp Cloud API. When agents record audio in Chatwoot, it is now
transcoded in the browser from WebM/Opus to OGG/Opus and sent with the
`voice: true` flag, so it appears as a native voice note bubble on
WhatsApp — not as a file/document attachment.
Closes#13283
**Key Changes:**
- Added `webmOpusToOgg.js` — a pure JS EBML parser + OGG page builder
that remuxes browser-recorded WebM/Opus audio into OGG/Opus entirely
client-side, with no server-side dependencies.
- Updated `AudioRecorder.vue` to use an explicit `mimeType` hint, proper
resource cleanup, and an `AUDIO_EXTENSION_MAP` for correct file
extensions.
- Renamed `mp3ConversionUtils.js` → `audioConversionUtils.js` and added
OGG conversion support via the new remuxer.
- Updated `ReplyBox.vue` to request OGG format for WhatsApp channels,
pass `isVoiceMessage` per-attachment, and handle recording errors with a
user-facing alert.
- Updated `MessageBuilder` to read the `is_voice_message` param and
persist it in attachment metadata.
- Updated `WhatsappCloudService` to:
- Normalize `audio/opus` → `audio/ogg` content type on ActiveStorage
blobs (works around Marcel gem re-detection).
- Send the `voice: true` flag when the attachment is a voice message
with `audio/ogg` content type.
- Use WhatsApp Cloud API `v24.0` for the attachment endpoint.
- Added `AUDIO_CONVERSION_FAILED` i18n key.
**How it works:**
1. The browser records audio as WebM/Opus (Chrome/Firefox default).
2. `audioConversionUtils.js` remuxes it to OGG/Opus using the pure-JS
`webmOpusToOgg` remuxer — no server transcoding needed.
3. The OGG file is uploaded with `is_voice_message: true` in the form
payload.
4. `MessageBuilder` persists `is_voice_message` in the attachment's
`meta` hash.
5. `WhatsappCloudService` normalizes the blob content type if needed,
then sends the attachment with `voice: true` so WhatsApp renders it as a
voice note.
## Type of change
- [X] New feature (non-breaking change which adds functionality)
## How Has This Been Tested?
1. Record a voice message in a WhatsApp Cloud conversation.
2. Verify the audio is transcoded to OGG (check file extension in the
attachment preview).
3. Verify the message arrives on WhatsApp as a voice note bubble (not a
document/file).
4. Send an image or document attachment and verify it still works as
before (no `voice` flag).
5. Send a regular (non-voice) audio file and verify it arrives without
the voice flag.
---------
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
Co-authored-by: Muhsin <12408980+muhsin-k@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Upgrades the frontend toolchain to Vite 6 and tidies up the build config
along the way. Behavior is unchanged for end users; this is dev/build
infra.
## What changed
- `vite` 5.4 → 6.4, `@vitejs/plugin-vue` → 5.2, `vite-plugin-ruby` → 5.2
(with matching `vite_rails`/`vite_ruby` gem bumps).
- Dropped the `vite-node` 2.0.1 pnpm override — no longer needed now
that vitest 3 runs on Vite 6 directly.
- Split the single `vite.config.ts` into:
- `vite.config.ts` (app), `vite.lib.config.ts` (SDK), `vite.shared.ts`
(aliases / Vue options), `vitest.config.ts` (tests).
- `pnpm build:sdk` now selects the SDK config explicitly instead of
branching on `BUILD_MODE=library`. SDK output path is unchanged
(`public/packs/js/sdk.js`).
No changes needed to Docker images, deployment scripts, or CI — Node 24
and pnpm 10 are already past Vite 6's floor, and the rake
`assets:precompile` hook still drives the SDK build via `pnpm`.
## How to test
- `pnpm dev` and verify the dashboard, widget, and survey routes load
and HMR works.
- Load a Chatwoot site widget on a test page and confirm `sdk.js` is
served and the widget mounts.
- `RAILS_ENV=production bundle exec rake assets:precompile` and confirm
`public/packs/js/sdk.js` plus the rest of the manifest are produced.
- `pnpm test` for the JS suite.
---------
Co-authored-by: Sivin Varghese <64252451+iamsivin@users.noreply.github.com>
Co-authored-by: Sony Mathew <2040199+sony-mathew@users.noreply.github.com>
When connecting an Instagram inbox during onboarding, the OAuth flow
used to drop users in inbox settings, breaking onboarding. The OAuth
start endpoint now accepts an optional `return_to=onboarding` hint,
carried tamper-proof inside the signed state (a claim on the Instagram
JWT), and the callback uses it to return the user to the onboarding
inbox-setup screen. Without the hint, behavior is unchanged.
This is the backend half only; the frontend that sends
`return_to=onboarding` ships separately.
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
## Description
Make conversation unread counts always available at runtime by removing
account feature checks from the API endpoint, unread-count listener,
notifier, and ActionCable broadcast path.
Update the dashboard to fetch sidebar unread counts for the active
account without checking FEATURE_FLAGS.CONVERSATION_UNREAD_COUNTS, and
remove the now-unused store clear action that only supported the
disabled state.
Keep the feature entry in config/features.yml to preserve flag bit
order, but mark it enabled and deprecated so fresh installs default to
the always-on behavior while feature-management UI hides it.
Leave existing installation default rows untouched; no migration is
included, so upgraded installs may still store the old flag value but
runtime behavior no longer depends on it.
Update specs around the new always-on contract and remove obsolete
disabled-feature assertions.
Fixes # CW-7237
## Type of change
Please delete options that are not relevant.
- [x] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality not to work as expected)
- [ ] This change requires a documentation update
## How Has This Been Tested?
Update specs around the new always-on contract and remove obsolete
disabled-feature assertions. Ran specs locally for the changes.
## Checklist:
- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [ ] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
Updates the Create New Message API documentation to explain how to send
file attachments with multipart form data.
Closes#10472Closes#12672
## Why
The endpoint already accepts attachment uploads, but the published API
docs only described the JSON request body. That made it unclear that
clients need to use multipart form data and send files through the
`attachments[]` field.
## What changed
- Adds `multipart/form-data` as a documented request body option for
Create New Message.
- Documents the `attachments` binary array and form encoding used by
`attachments[]`.
- Adds a multipart cURL request example for a message with an
attachment.
- Regenerates the Swagger JSON artifacts.
## Screenshots
Create New Message API docs with the multipart cURL sample and
`multipart/form-data` request sample selected:
<img width="1702" height="1083"
alt="create-message-attachment-docs-multipart"
src="https://github.com/user-attachments/assets/5fef9082-6eb9-494d-91d1-8dc0b7880212"
/>
## How to test
1. Open `/swagger`.
2. Navigate to Application -> Messages -> Create New Message.
3. Confirm the endpoint documents both `application/json` and
`multipart/form-data`, including the attachment payload schema.
When connecting a Gmail or Outlook inbox during onboarding, the OAuth
flow used to drop users in inbox settings, breaking onboarding. The
OAuth start endpoint now accepts an optional `return_to=onboarding`
hint, carried tamper-proof inside the signed `state`, and the callback
uses it to return the user to the onboarding inbox-setup screen. Without
the hint, behavior is unchanged.
This is the backend half only; the frontend that sends
`return_to=onboarding` ships separately.
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
