Integrate.io Release Notes & Product Updates

July's release cycle put AI assistance at the center of the package designer, covering pipeline building, field mapping, and expression writing. Alongside that, the editor gained safer handling of in-progress work, workflow orchestration went a level deeper, and destination coverage expanded.

Helm: Prompt to Pipeline

Helm is your copilot in Integrate.io. It builds pipelines, answers questions, and manages your packages.

• Build a pipeline from a prompt, or start from a ready-to-run package and adapt it
• Answer questions about how a component behaves, with a link to the source
• Manage packages through edits that land in a draft, not the live version

Helm designs and builds; the Integrate.io platform executes. A pipeline that ships stays deterministic and runs the same way every time. Nothing becomes production until you review and approve.

Helm Ask is on for every account, Helm Build is enabled per account by our team, and the documentation has the full picture.

Smart Map and Smart Transform Take the Slowest Step Out of Pipeline Delivery

Smart Map removes field-by-field setup. Point the platform at your destination schema and it automaps your source data to the matched schema output, including renames and type conversions. Smart Transform works at the field level: describe the transformation in a prompt or give a before and after example, and it generates the Integrate.io transformation expression for you. Suggestions are validated against your own data before you accept them, so what you review is what runs. Both are available on request, and the documentation covers what they can do.

Package Drafts and Autosave Keep In-Progress Edits Safe

Losing package edits to a closed tab or a dropped connection used to be a real risk, since nothing was stored until you explicitly saved a version. The designer now autosaves your work to a private per-user draft after you stop editing, and again when you leave the editor. Reopening a package that has a draft asks whether to resume the draft or open the last saved version, and a Discard Draft button in the toolbar throws the draft away when you want to start clean. Draft autosaves stay out of version history, so only explicit saves create a version. Full setup details are in the documentation.

The Run Workflow Task Runs Workflows Inside Workflows

Reusing one workflow as a step inside another previously meant duplicating its tasks by hand and keeping both copies in sync. The Run Workflow task executes another workflow as a sub-run of the current one, so a long orchestration can be split into smaller workflows that also run on their own. Read the documentation for the nesting limit and variable scoping rules.

New Connectors

• Salesforce REST API Destination: A new Salesforce destination built on the REST API supports insert, upsert, update, and delete across throttled single-record and batched write modes, and produces per-record success and failure CSV outputs. It works with both standard and customer-managed OAuth Salesforce connections, giving teams an alternative to the Bulk API for smaller or latency-sensitive writes (documentation).
• Databricks Destination: Pipelines can now write to Databricks, with staging and load handled for you so output lands in your tables without a manual copy step. Teams running on the lakehouse can send ETL output to Databricks directly instead of routing it through an intermediate warehouse first (documentation).
• Trustpilot Source: A native Trustpilot source pulls reviews, review categories, and private reviews for a business unit, with incremental load supported. Review data can now be joined against the rest of a customer dataset without maintaining a custom API pull (documentation).

New Features

• Live Preview in the Expression Editor: The expression editor now runs the current expression against five real sample rows from the previous component and shows each input value next to its result. Expressions can be checked while you write them instead of after a job fails.
• Field Data Types in the Select Component: The Select component's field list now shows a data type for every output field, and the expression editor shows the return type for most functions, fields, and variables. Where a type cannot be inferred statically, running a preview fills the column in.
• Separate Read and Write Access for AI Tools: Helm and MCP capabilities are now split into distinct read and write tool sets, so inspection, previews, and searches can be opened up broadly while package edits, schedules, and job runs stay restricted. Write access is closed by default, and our team can open it up for an account on request.
• Package Group to Connection Group Binding: A package group can now be bound to specific connection groups, restricting its packages to connections from those groups. Admins get a direct way to keep one team's pipelines from reaching another team's credentials.
• AddToDate and ToDateOrNull Functions: Two new expression editor functions cover date arithmetic and parsing that returns null rather than failing on an unparseable value. Date handling that previously needed nested conditionals now fits in a single call.

Improvements

• SFTP connections now support private key plus password authentication, covering servers that require a multi-factor login.
• Job logs, the logs API, and AI error summaries mask connection credentials, and account-level global secrets are masked in job and workflow variable views.
• Chatmeter, Corrigo, and GoTo sources now paginate through results, so large accounts return their full dataset instead of a first page.
• The Email Report destination accepts variables in its subject and body.
• HubSpot destination has a new Clear when empty flag that clears a HubSpot property when the incoming value is null.
• File and cloud storage destinations can skip writing an output file when a run returns zero rows, an opt-in setting for CSV and JSON.
• The Google Sheets source imports the full sheet schema without manual selection, and can fail a job when the sheet contains no data rows instead of finishing with nothing loaded.
• The Google Ads source adds ConversionsByConversionDate and ConversionsValueByConversionDate to its available metric fields.
• The REST API destination supports a sleep interval between batched requests, giving rate-limited endpoints room between calls.
• Custom role management adds a filter on the roles list and a form for assigning users to a role directly, and the packages and connections lists gained a Show ungrouped only switch.

Fixes

• Package variables are no longer removed when a variables save and a package save run at the same time.
• Package validation succeeds when ExecuteSql or ExecuteSqlDateTime is used in a package variable.
• Draft variables are now executed during schema preview calls.
• The Select component opens correctly when an integer is used as a package variable.
• Source fields with commas in their names, such as long Google Sheets headers, now carry their values into downstream components instead of previewing as null.
• File Storage jobs run correctly when a filename contains parentheses or the source path contains spaces.
• Packages whose variable substitutions produce long file paths no longer crash mid-run.
• SFTP and FTPS sources using a bare wildcard resolve the same file in preview and at job runtime.
• The DB Lookup component's + New button now opens the connection form.
• The Salesforce destination writes error and success record files to shared Google Drive directories when a directory ID is specified.
• The AI error summary reports Salesforce write failures correctly instead of describing them as filesystem handler errors.
• Packages inside a package group with connection group restrictions now save correctly.
• Global secret variables evaluate correctly in the cases that previously failed.
• Invitations work for users assigned a custom role, and the user management page is restricted to administrators.
• The owner role shows its label and filter on the user list and is assigned correctly on the roles view.

Performance Enhancements

• The Google Ads source can split a long date range into parallel chunks, cutting pull times on wide reporting windows. Set _ADWORDS_API_DATE_SPLIT_ENABLED to true in your package variables to switch it on, with _ADWORDS_API_DATE_SPLIT_CHUNK_DAYS and _ADWORDS_API_DATE_SPLIT_MAX_CHUNKS controlling the split.
• The Google Sheets source reads large sheets in batches rather than loading a whole sheet in a single pass, clearing the out-of-memory, read-timeout, and missing-row failures that wide or long sheets used to hit.

Give the new AI tooling a run this month and tell us where it saves you time and where it still needs work.

July's release cycle extended destination coverage to the lakehouse and added a widely requested support-platform source. The rest of the cycle went into initial-sync reliability and giving teams a way to verify their alerting before they need it.

The Databricks Destination Lands Replicated Data in Delta Tables

Teams running on the lakehouse previously had to route CDC output through an intermediate warehouse before it reached Databricks. Pipelines can now replicate into Delta tables directly, with staging and load handled for you so resyncs, CDC updates, and scheduled runs behave the same way they do on every other warehouse destination. Type conversion is mapped across ten source types, including MySQL, Postgres, SQL Server, Oracle, DB2, Salesforce, BigQuery, and file storage, so high-precision decimals, Postgres complex types, SQL Server TIME, and JSON survive the load intact. Setup runs through a guided connection flow in the app, and AWS is supported today with Azure and GCP planned.

New Connectors

• Intercom Source: Intercom is now a supported source via OAuth with eleven streams covering contacts, companies, conversations, conversation parts, tickets, admins, teams, tags, segments, data attributes, and articles. Support and product teams can replicate their full Intercom history into the warehouse alongside the rest of their customer data.

New Features

• Slack Notification Hook: Pipeline run notifications can now go to Slack through a one-click OAuth connection in the Hooks UI. Teams get failure and completion alerts in the channel they already watch instead of only by email.
• Test Hook: A Test Hook action in the Hooks menu and in Settings sends a real ping to a configured Slack channel or email recipient list. Hook configuration can be verified at setup time rather than on the next failed run.
• Expanded Stripe Streams: Refunds, payment methods, and dispute balance transactions are now available as Stripe streams. Finance teams can reconcile chargebacks and refunds in the warehouse without maintaining a separate extract.
• Deleted Ad Records in Facebook Ads and TikTok Ads: Deleted ad rows are now included by default on supported streams, with a one-time historical backlog arriving on the first sync. Contact our team if you would rather exclude them.

Improvements

• Editing a hook preserves the fields you did not change, and an invalid configuration is rejected with a clear error instead of being saved.
• Databricks resyncs drop and recreate the destination table by default, with an opt-out for teams that manage table lifecycle themselves.
• The Intercom connector recovers on its own when a sync cannot resume from its last position, instead of stopping until someone intervenes.
• Intercom data attributes now use a composite primary key, so tables created before this change, including data_attributes and tickets, need a drop and resync.

Fixes

• Initial syncs no longer produce duplicate rows on affected pipelines, though a small number of tables that were mid-sync may need an explicit resync.
• DB2 to BigQuery initial syncs complete instead of hanging indefinitely.
• Pipelines that stop with an error fire the Pipeline Errored notification instead of Pipeline Stopped.
• The Bing Ads campaign stream includes the ad type that was previously missing.
• Intercom syncs complete for organizations with more than 10,000 records and recover from expired sessions.

Performance Enhancements

• Initial syncs to BigQuery run in parallel with safer retry handling, cutting the time it takes to load a large source.
• Out-of-memory failures during large initial syncs have been resolved, so a parallel sync finishes rather than ending partway through.

Let us know which sources you want us to cover next.

June's release cycle opened up finer-grained account access control alongside new native building blocks for calling APIs, enriching rows, and delivering pipeline output by email. It also introduced a documented way for AI assistants to work directly with an account, rounded out by hardening across Salesforce and other components and a large batch of editor and connector fixes.

Custom Roles Bring Precise, Scoped Access
Predefined roles like Owner, Admin, and Editor don't always match how a growing team is actually organized, either handing out more access than a group needs or leaving no way to grant a narrower set of permissions. Custom roles let account admins build their own permission sets and scope them to specific package groups or connection groups, so a team working on one set of pipelines never sees another team's packages, jobs, or connections. Changes take effect immediately for every member assigned to them, so access can be tightened or extended without re-inviting anyone. See the documentation for a full walkthrough of scopes and permissions.

The API Call Component Handles Lookup via APIs
Calling an API mid-pipeline no longer means writing and debugging a Curl() expression by hand. The new native API Call component handles URL generation, headers, and authentication through a UI-first design, and parses the response with an auto-detect JSONPath in the same step, so pipelines that depend on live API data are faster to build and easier to maintain. Full setup details are in the documentation.

The Lookup Component Makes Row Enrichment Easier
Enriching rows with reference data used to mean reaching for a full join, even for a small lookup table. The new Lookup transformation component handles that case directly, pulling in columns from a database table or query with far less setup, and now supports fuzzy matching for the common case where reference keys do not line up exactly, so teams spend less time wiring up enrichment logic and more time on the transformation that actually matters. Read the documentation to learn more.

The Integrate.io MCP Server Connects AI Assistants to Your Account
Checking on a pipeline's status, diagnosing a failed job, or previewing source data before building a transformation usually means switching over to the dashboard and clicking through several screens. The new MCP Server exposes a documented set of tools that let AI assistants like Claude Desktop, Claude Code, and Cursor connect directly to an account and answer those questions on the spot. Beyond read-only inspection, a set of guarded tools can also clone a package, edit components, manage a schedule, or run and retry a job, so routine account work can happen from inside the assistant a team already uses. Get set up with the documentation.

The Email Report Destination Delivers Output Straight to Inboxes
Getting pipeline output into someone's inbox used to mean landing a file in storage first, then routing it onward by hand. The new Email Report destination component sends data straight to an email address, either as body text or as a CSV attachment, cutting that storage hop out entirely and making reporting workflows easier to set up and maintain. Check out the documentation for configuration options.

New Features

• Connection Groups: Connections can now be organized into groups the same way packages are organized into Package Groups, making it easier to navigate accounts with a large number of connections.
• Dynamic Header Detection for File Storage: CSV and Excel sources can now detect the header row automatically by matching it against expected schema column names, instead of always assuming row one. This fixes a recurring problem for multi-file ingestion where header position shifted between files, including files where the header row contains a hyphen.
• Google Drive Source Component: A native Google Drive source is now available with both OAuth and service account authentication, letting users pull files directly from Drive into a pipeline without routing through a generic file storage connector.
• Global Secret Variables: Account-level secret variables are now available, letting teams define a credential or token once and reference it across packages instead of duplicating it in each one.
• Incremental Load Last Successful Run Option: Source components can now set incremental load to use the package's last successful run timestamp automatically, instead of manually wiring up the underlying system variable.
• Dummy Source Component: A new source component supports packages that take their data entirely from package variables or from upstream transformation components like Python Transformation or API Call, without requiring a placeholder source connection.
• New Expression Editor Functions: ExcelSerialToDate, DateToExcelSerial, and FUZZY_MATCH are now available in the expression editor, covering Excel date conversion and string similarity scoring (0-100) without custom logic.

Improvements

• Salesforce proxy is now enabled by default for all accounts, routing Salesforce traffic through Integrate.io's IP range.
• Salesforce destination error and success record files can now be uploaded to GoogleDrive.
• NormalizePhone now accepts an optional region argument and an on_invalid mode, letting a pipeline fall back to a best-effort E.164 number or pass the original value through instead of dropping invalid numbers to null.
• Expression editor autocomplete now automatically wraps expressions in parentheses, advances the cursor to the next argument when a function takes multiple arguments, and triggers without requiring a leading space.
• File Storage Source now automatically appends _1, _2, and so on to duplicate column names detected during schema import.
• GoogleSheet destination now supports specifying a custom range and an optional header control when using overwrite mode.
• Improved search on the Packages and Jobs lists. You can now search by partial words, and it matches names that include special characters.
• Package Designer notes can now be assigned a color, and components support an optional description field for documentation within the canvas.
• Database Connection forms now validate colon and semicolon characters correctly, with validation messages positioned where users expect them.
• Facebook Ads connector now uses comscore_market in place of the deprecated dma field, keeping reporting data flowing as the upstream field is retired.

Fixes

• Salesforce Destination now delivers the success-records file even when a job partially fails, instead of only the failed-records file.
• Salesforce SOAP Destination now produces a CSV file even when maximum errors is set to 0, and checks failed records across all batches against the configured limit instead of only per batch.
• Google Sheets destination overwrite mode no longer fails with a "Range exceeds grid limits" error.
• SFTP connections with an Ed25519 SSH key authenticate correctly instead of failing the job.
• REST API source parses nested JSON objects correctly when Unwrap extracted string values is selected.
• Package Designer no longer places a newly added note or component outside the current viewport after scrolling.
• Package Designer no longer shows a white background while the browser or assistant window is resized.
• Package validation no longer fails for packages referencing newly added, unsaved variables.
• Database CDC component now generates correct SQL for PostgreSQL databases with a dedicated schema name.
• Snowflake CDC component is available again as its own component instead of only nested under Database Source.
• Workflows now use the current variables and secrets from the related package, and correctly apply overwritten variables in the Run Package component.
• Rolling a package back to an older version keeps it in its original Package Group.
• Export Workflow package now reflects a renamed package correctly inside the Run Package component.
• Join component no longer breaks the component preview when swapping inputs.
• Python Transformation's Run Code button no longer hangs indefinitely.
• Excel date values ingest correctly in both the previewer and job runs.
• Hubspot sources return *_id columns typed as String.
• File Uploader job runs are no longer failing on some regions.

Take a look through what shipped this cycle and tell us which of these changes make the biggest difference for your team.

June's release cycle focused on expanding destination coverage with three new connectors and opening up programmatic pipeline management through a public API. Alongside that, sync reliability and security posture were hardened across MySQL, PostgreSQL, and Snowflake.

The Public API and YAML Import/Export Enable Programmatic Pipeline Management
Pipelines can now be managed programmatically through a new public API, with support for exporting and importing pipeline definitions as YAML. Teams can version pipeline configurations alongside application code, replicate pipelines across environments, and manage pipelines from their own tooling instead of relying solely on the dashboard. See the documentation for endpoint reference and authentication setup.

PostgreSQL, MySQL, and SQL Server Destinations Round Out Database Coverage
These Databases are now available as destinations, rounding out the platform's database destination coverage alongside existing targets. Teams running on these databases can land CDC data directly into them without staging through an intermediate warehouse first.

New Features

• Google Ads Expanded Performance Report Fields: CAMPAIGN, AD, ADGROUP, BUDGET, and KEYWORDS performance reports now expose significantly more fields per stream, aligning their schemas with the equivalent entity-stream reports and giving analysts access to metrics that previously required a separate report.
• MySQL and BigQuery Schema Change Hooks: MySQL and BigQuery sources now emit NEW_COLUMN and NEW_TABLE hook events, letting teams trigger downstream automation when source schemas change instead of polling for new fields or tables.
• Hubspot Custom Object Support: Hubspot sources can now sync custom objects in addition to standard objects, giving teams with custom Hubspot schemas full pipeline coverage instead of being limited to default object types.

Improvements

• PostgreSQL destination's graceful termination handling has been improved, and credential inputs are now available directly in the UI setup instead of requiring manual configuration.
• Snowflake destination role requirements have been hardened to least privilege, dropping the OWNERSHIP and schema MODIFY privileges that were previously required to sync data.
• MySQL source now supports per-table error isolation, so an issue with one table no longer halts replication for the rest of the source, and an optional GTID rewind override is available when replication state is ahead of the source.
• PostgreSQL destination now supports an override to make the heartbeat table optional, accommodating deployments that restrict heartbeat table creation.
• Parallel chunk workers on sources writing to BigQuery destinations have been disabled by default to avoid tripping BigQuery rate limits during sync.
• Hubspot connector now automatically syncs new columns added to custom objects, extending the custom object support shipped earlier this cycle.

Fixes

• PostgreSQL destination no longer produces ghost rows during a full resync, and no longer requires table ownership or CREATE privileges to sync.
• Non-PostgreSQL destinations no longer fail from a regression tied to a null file compression configuration value.
• Snowflake connector's test connection setup query helper no longer returns undefined values.
• MySQL connector no longer mismaps uint columns, JSON columns during parallel sync, or datetime values nested inside JSON during parallel sync.

We'd love to hear how these updates land in your day-to-day syncs, so don't hesitate to reach out with feedback.

• INITCAP(): A new expression editor function capitalizes the first letter of each word in a string and lowercases the rest, covering a common text normalization step that previously required custom transformations.
• NormalizePhone(): Standardizes phone numbers into E.164 format, handling a wide range of input formats and providing configurable behavior for invalid numbers - removing the need for manual cleaning logic in pipelines that ingest contact data.
• NormalizeDate(): Accepts datetime inputs and returns the date portion, validating against multiple formats so users can reliably extract dates from mixed-format fields without writing custom expressions.
• NormalizeLanguage(): Normalizes language codes in the expression editor, expanding the set of standardization tools available when working with localization or audience data.
• Excel Sheet Selection for Large File Ingestion: When the "ingest large Excel file" option is enabled, users can now specify which sheet to load, supporting workbooks where the target data is not on the default sheet.
• Inline SFTP Key per Connection: Each SFTP connection can now carry its own Ed25519 key pair inline, giving users connection-level key management without relying on shared credentials.
• Create New Option in Cluster and Connection Pickers: Dropdowns for cluster and connection selection in the Job form and component editors now include a "Create new" option, letting users add a resource without leaving the current context.
• Package Variable Persistence: Saving package variables now immediately persists them to the database and triggers draft validation, so newly added variables are recognized in the package without requiring an additional save cycle.

• The Mouseflow connector has been updated to move the website ID input from the connection form into the source component, aligning configuration with where it is most relevant.
• Gmail source now extracts full email bodies, replacing the previous snippet-only behavior and supporting a dedicated "Email body" field option for ingesting content without attachments.
• The Snowflake connection form has been updated from Account identifier to Account locator, reflecting the current Snowflake authentication model and reducing configuration errors.
• Azure REST API pagination is now supported for the management.azure.com endpoint, enabling reliable ingestion of paginated Azure resource data.
• AI error summaries now anchor on the deepest root cause in stack traces rather than the first exception, producing cleaner, more actionable explanations for Salesforce timeouts, S3 access errors, schema mismatches, and cluster failures.
• The Package Designer now opens the relevant component editor when a user clicks on an error in the Package Errors panel, reducing the steps needed to navigate from an error to the fix.

• Component previewer row ordering now matches schema preview ordering.
• Sort and Window components now return rows in the correct sequence during preview.
• Window component no longer adds "stitched::" prefixes to field names in the component previewer.
• Limit component previewer now works correctly.
• Window and Union component validation no longer incorrectly disables save and preview buttons after input updates.
• Universal OAuth connector reconnection for Google OAuth now works correctly - missing parameters that previously broke the request have been restored.
• Facebook Ads destination now shows the correct operation type options for the Customers object when reopening a saved configuration.
• AI Assistant popup no longer appears behind the expression editor and other modals.
• Expression Console search in the functions popup now returns correct results.
• Multi-line CSV parsing in Cloud Storage and SharePoint sources now works correctly when a text qualifier (" or ') is set.
• Google Sheets source headers containing commas or newlines no longer shift downstream columns, and duplicate-header columns now render their values correctly.
• Renaming a workspace no longer incorrectly updates the Last Modified date on every package inside it.
• Salesforce proxy traffic now routes through the correct region for proxy-enabled accounts.

• The hook filter editor now uses lazy loading for package and workspace lists, reducing memory usage and keeping the UI responsive in large accounts.
• JSON extraction functions have been optimized to handle a wider range of data types more efficiently, reducing processing overhead for pipelines that work with complex nested structures.