Performance & Reliability
Contacts and companies now appear much faster after import. We overhauled the way data is indexed in the background — processing now happens in batches instead of one record at a time, resulting in roughly a 50× speed improvement. In practice this means:
Contacts imported via CSV or LinkedIn search appear in the CRM within minutes rather than hours
Dashboard metrics (connection requests sent, tasks completed, etc.) update much more reliably after actions run
Imports no longer get stuck in the queue for extended periods
Infrastructure upgrade for the first cluster. We migrated the first server cluster to new hardware. All data was preserved and the system is running significantly faster across the board. If you previously experienced lag, slow task execution, or delayed automation actions, this should be noticeably improved.
LinkedIn & Connection Requests
More reliable connection request limit handling. When a sender hits LinkedIn's weekly connection limit, the system now double-checks before setting the limit — matching the behavior you'd see in LinkedIn's own browser interface (where the first attempt sometimes shows a limit warning but the second goes through). This reduces cases where limits are applied too early.
Updated connection limit message. The weekly limit warning has been updated to better reflect how LinkedIn actually applies limits — which can reset daily in some cases, not always weekly. The new message reads:
"LinkedIn connection limit reached. Automatic checks run every few hours, and sending will resume once the limit is lifted."
InMail reliability improvements. Several fixes were made to how InMail tasks handle limits and errors:
Daily and monthly InMail limits are now detected more accurately, with retries scheduled for the next day
Generic LinkedIn errors (HTTP 400) no longer trigger a sender limit
Imports
LinkedIn Filters are now required when creating a Find Contacts import. Previously it was possible to create a Find Contacts import without any LinkedIn filters, which caused browser errors during automation execution. The form now validates that at least one filter is set before the import can be saved.
Company import fixes. Two bugs affecting company imports have been resolved:
When the Add tags to duplicate accounts option was enabled, tags were not being applied to companies flagged as duplicates. This is now fixed.
When the Move to list if already saved option was enabled, duplicate companies were not being added to the specified list. Also fixed.
"Everyday" repeat period added for Post Engagement imports. You can now set Post Engagement imports to repeat daily, in addition to the existing weekly/monthly options.
CRM & Contacts
"Automation finished" activity now shows the historic pipeline stage. The activity log entry created when a contact completes an automation ("Automation X finished. Current lead status: …") now shows the status the contact had at the time of completion, not their current status. This makes the history more accurate for contacts whose status changed after the automation ended. Applies to new activities only — existing entries are not retroactively updated.
Email & Mailboxes
Email sync error fixed. An internal error during email body synchronization ("The selected status is invalid") could occasionally stop a mailbox from syncing. This has been resolved.
Custom Tracking Domain now accepts domains with hyphens. Domains like track.mail-example.com were being rejected by the validation check when adding a Custom Tracking Domain to a mailbox. This is now fixed.
Search Page Import Limit resets automatically. Previously, the Search Page Import Limit for sender profiles would only reset when the configured limit was reached. It now also resets automatically if it hasn't been updated in 12+ hours (checked every 6 hours). This fixes cases where the limit was not resetting daily as expected, causing slower-than-expected import speeds.
API
Upsert no longer clears unspecified fields. When updating a lead via the API upsert method, fields not included in the request body were being set to null — for example, if first_name wasn't passed, it would be wiped. This has been fixed: only the fields explicitly included in the request are updated.
AI Features
Gemini template generation is working again. An error in AI template generation when using the Gemini provider (caused by a Google API change) has been resolved.
Comment templates have been temporarily removed. The ability to create AI templates for LinkedIn post comments and select them in the automation builder has been removed while we rework this feature. The current implementation had reliability issues and the output quality wasn't consistent enough.