Connectors Changelog

Updated debug logging so failed requests report the error message instead of an empty object

Updated the New and Updated CRM Leads, New and Updated CRM Contacts, and New and Updated Books Contacts polling triggers to filter records server-side by their last modified time, so each poll returns only records changed since the previous poll instead of the entire dataset, reducing API-credit consumption

Reworked the New and Updated Records and New and Updated Custom Records triggers for large data volumes and correct filtering:

• Added an optional Look-back Date input across all polling triggers for performing an initial sync of records. The initial sync backfills every record created on or after the specified date, seeding each once and ignoring the field and visibility filters; later recurrences are unaffected. Leave it empty to start from the first recurrence with no backfill
• Added opt-in batching across all polling triggers, dispatching each changed record individually or in configured batches so a large backlog drains within one recurrence; enabling it changes the shape a downstream step receives
• Added support for object types holding more than the 10,000 records a single HubSpot search returns, draining the remainder on later recurrences. A larger result set previously failed
• Updated both triggers to always return both createdRecords and updatedRecords. A disabled Show New Records or Show Updated Records now yields an empty array where the key was previously absent
• Updated Search Properties filters and filterGroups to restrict results instead of widening them. They were previously OR-combined with the trigger's date window, so configured filters returned matching records regardless of date
• Updated Search Properties sorts to be ignored by these triggers; they now sort ascending by the object's creation property during the initial sync and its last-modified property on later recurrences, so a recurrence can resume where it left off
• Fixed the New and Updated Custom Records trigger retrieving every record of the custom object type on each recurrence instead of only the changed ones. It now filters and sorts on hs_lastmodifieddate, so the custom object type must expose that property

Updated the New and Updated Records polling trigger:

• Updated the trigger to carry unreported records forward to the next recurrence, so a backlog larger than Max Records To Fetch is no longer dropped and a bulk load stamping many records with one timestamp still drains; Max Records To Fetch now caps how many records are retrieved at a time rather than how many the trigger will ever report
• Added an optional Look-back Date input for performing an initial sync of records. The initial sync backfills every record created on or after the specified date, seeding each once and ignoring the field and visibility filters; recurrences after the sync are unaffected. Leave it empty to start from the first recurrence with no backfill
• Added opt-in batching to the New and Updated Records trigger, dispatching each changed record individually or in configured batches so large backlogs drain in one recurrence; enabling it changes the shape a downstream step receives
• Updated deleted-record reporting so a trigger that has not run in more than 15 days continues instead of failing, since Salesforce serves no more than 15 days of deletions, and so deletions at the edge of a window are reported on the next recurrence rather than skipped

Restructured list action inputs into structured objects and improved paging reliability:

• Grouped the page and page-size inputs on List Accounts and List Properties into Pagination; Fetch All stays a top-level toggle
• Fixed Fetch All on List Properties, and the properties dropdown, so the account filter is applied to every page rather than only the first
• Fixed list actions, data sources, and the New and Updated Records trigger to return an empty result instead of erroring when no records match
• Added output schemas to the List Accounts, List Properties, Get Property, Run Report, and Send Measurement Protocol Events actions for improved field mapping during configuration
• Added inline action calling support across all actions for improved example output during configuration

• Corrected the Limit range on the List Models action from 1-3 to 1-1000 to match Anthropic's documented maximum; Fetch All now pages at 1000 records instead of 3
• Updated the default model from claude-3-5-sonnet-latest to claude-sonnet-4-6 across all actions
• Grouped pagination inputs on the List Models action into a Pagination structured object, with Fetch All remaining a top-level toggle
• Added output schemas to the Chat, Count Tokens, Get Model, and List Models actions for improved field mapping during configuration
• Added inline action calling support across all actions for improved example output during configuration

Enhanced action configuration and connection security:

• Grouped the paging and query-control inputs on List Devices, List Groups, and List Packages into the Pagination and Filters structured objects; Fetch All remains a top-level toggle
• Fixed List Devices to send configured custom query parameters, which were previously discarded
• Changed the API Key connection field to a masked password field
• Added output schemas to the Create Deployment, Get Device, List Devices, List Groups, Get Package, and List Packages actions for improved field mapping during configuration
• Added inline action calling support across all actions for improved example output during configuration

Improved action configuration and connection setup guidance:

• Added setup steps and vendor documentation links to the OAuth 2.0 and OAuth 2.0 Client Credentials connection documentation, including the callback URL required to complete the OAuth 2.0 authorization flow
• Added output schemas to 53 actions for improved field mapping during configuration
• Added inline action calling support across all actions for improved example output during configuration
• List actions (List Assets, Query Assets, List Automations, List Campaigns, List Categories, List Data Extensions, List Journeys, List Email Definitions, List SMS Definitions) group their pagination controls into Pagination; Fetch All stays a top-level toggle
• Create Campaign groups its optional fields into Additional Fields, keeping Campaign Name at the top level
• Corrected the Message Key description on Send Email, which described the input as the email definition identifier rather than the unique key that identifies an individual send, and the Parent Category ID description on Create Category, which described a required input as optional
• Renamed the Extra Body input on Update Automation to Automation Properties

Enhanced component capabilities:

• Updated the Select Topic data source to read the cluster address from the Brokers input so topic dropdowns resolve (breaking: the data source's Broker input was removed, so previously configured values must be re-entered)
• Grouped the Session Timeout and Heartbeat Interval inputs on the Kafka Consumer trigger into a Session Timing structured object
• Added output schemas across all actions for improved field mapping during configuration
• Added inline action calling support across all actions for improved example output during configuration

Restructured action inputs into structured objects for an improved user experience.

• List actions (List Users, List Managed Devices, List Mobile Apps, List Managed Apps, and List Detected Apps) group their OData inputs into Pagination and Filters; Fetch All stays a top-level toggle
• List Directory Audits groups its OData inputs into Pagination and Filters; Fetch All stays a top-level toggle
• List Domains groups its OData inputs into Pagination and Filters
• List Groups, List Group Members, and List Software Update Status Summary group their OData query inputs into Filters
• Update User groups First Name and Last Name into a Name structured object
• Update Software Update Status Summary groups device-level compliance counts into Device Counts and user-level compliance counts into User Counts
• Create Managed App, Update Managed App, Create Group, and Update Group group their secondary metadata inputs into Additional Fields
• Renamed the Resource Trigger trigger to Resource Change