Changelog

New ✨

AI Skills library
Browse, search, and filter core and catalog skills from the new skills library at https://app.coupler.io/app/ai_skills.

Install or uninstall skills, create and edit your own custom skills, and view skill detail pages — all with role-based permissions. Skills are also selectable directly from the slash menu in the AI agent composer, and can be resolved via MCP for installed skills.

Amazon Seller Central source
Connect and import your Amazon Seller Central data (non-sensitive scopes for now, with sensitive-scope approval in progress).

Pipedrive: new source on the v2 API 
Pipedrive is now available as a new source built on Pipedrive's v2 API. This is a separate source from the legacy version — existing data flows on legacy Pipedrive keep running as-is, but to move to the new source you'll need to recreate your data flows against it. Find your legacy flows on the data flows page by searching "legacy" with the Pipedrive source filter. Dashboard and data set templates have been updated to use the new source.

Invitations now require acceptance
When you invite someone to a workspace or organization, they get an email with an "Accept invitation" link instead of getting instant access. Invited users show as "Pending" until they accept, and only "Active" users count toward your billing limit. Invitation links are valid for 14 days and can be revoked before they're accepted.

Improved 👍

Zoho Books: two new data entities
General Ledger and Chart of Accounts are now available as Zoho Books data entities.

Instagram Insights: new metric in lifetime report
Added "Total views" to the "Post: performance lifetime report."

Join/Append/Aggregation wizards: full column names on hover
Long column names now truncate in these wizards, with the full name shown on hover.

Fixed 🐛

Integration fixes
Resolved issues across several integrations: Freshdesk (Tickets imports capped at 200 rows), Adobe Commerce/Magento (Orders incremental sync not saving state), Freshbooks (missing pagination parameter for Invoices), Teamwork (missing records from pagination gaps), RD Station Marketing (timestamp timezone errors), LinkedIn Company Pages (report period errors), Brevo (crash importing email campaigns with an expiration date), Facebook Ads (crash on object breakdowns), Airtable (unsupported field types failing the destination load), Retail Express (Maropost) (fixed timeouts during data load).

Destinations: failed run error page
A failed destination run now shows the actual error instead of redirecting to datasets.

AI agent: Cursor connector setup
Fixed an OAuth error that prevented the Cursor connector from completing setup.

Google Sheets add-on: Helpscout beacon
Fixed the Helpscout support beacon not fitting the screen.

New 🌟

AI agent: schedule your dataflows Ask the agent to set a dataflow's refresh schedule — enable or disable it, set the interval, choose days, start time, and timezone.

Oracle NetSuite source Connect to NetSuite as a data source with token-based authentication. Pull data using dynamically loaded objects (like Account) with a start-date filter, or write custom SuiteQL queries. Invalid SuiteQL now returns a clear error instead of failing silently. 

Visit to learn how to setup & what data available ➡️ https://www.coupler.io/sources/netsuite-connector

HubSpot: Projects entity HubSpot Projects can now be imported as a data source. If you're on existing HubSpot credentials, you may be prompted to reconnect to grant the new required scopes.

Improved 💪

Meta Ads: Comscore Markets breakdown Facebook and Instagram Ads reports now offer a Comscore Markets breakdown, replacing the DMA breakdown that Meta retired.

BigQuery source: accurate types and faster imports Schema is now built directly from your query definition — columns appear in query order, dates and numbers are typed correctly, nested structs flatten to readable field names, and large imports run faster.

Brevo: custom contact attributes Custom attributes on the Contacts entity are now included in your data.

AI agent: smoother wizard integration When the agent adds or updates a source or destination, changes now appear in the wizard instantly without a refresh. Setting up a destination, the agent now prompts you to pick (or auto-selects) the right dataset to share. And AI actions like Context and Generate Schema are disabled with a clear explanation when a dataset is marked "Exclude from AI agent."

Tidier run history Long source names in your run history are now truncated with an ellipsis instead of causing horizontal scroll.

Fixed 🚧

Klaviyo: sources from templates Klaviyo sources created from a template no longer break due to a mismatched metric ID.

HubSpot: readable names for custom objects Custom objects now return readable pipeline and stage names, matching how default objects like Deals already worked.

Appfigures: reports now respect your date range Reports correctly break down by date and group-by selection instead of collapsing into a single all-time total row.

Xero: Budget Summary report In Date-by-row format, the Budget Summary report now includes the account name field and no longer returns total-only rows with no account.

Twitter Ads: fixed analytics crash Resolved a crash that occurred while polling Twitter Ads async analytics job status.

Dashboards: layout preserved on duplicate Duplicating a dashboard no longer breaks its layout.

Google Sheets: correct destination naming Setting up a data flow from a Google Sheets template now names the destination after your actual copied spreadsheet, instead of the template's placeholder name.

Improved

Meta integrations updated to API v25
All 6 Meta sources — Facebook Page Insights, Facebook Public Data, Instagram Insights, Instagram Public Data, Facebook Ads, and Instagram Ads — are now on the latest API. Facebook and Instagram Ads also include new error_* fields for async report failures, giving more detail when a report run doesn't complete successfully.

Facebook Page Insights: deprecated metrics removed
Metrics and reports deprecated by Meta have been removed from Facebook Page Insights. This affects the Page performance insights, Post lifetime performance, Post reels lifetime performance, and Post video daily performance reports. See the full list of removed metrics.

Facebook & Instagram Ads: updated attribution windows
Facebook Ads now supports additional manual attribution window options: incrementality, engagement-based, and first/all conversions. 

BigQuery source: large dataset support
The BigQuery source performance improved to handle big volume data.

BigQuery destination: table descriptions preserved for Gemini
Coupler.io metadata is now written to BigQuery table labels instead of table descriptions. This keeps your descriptions intact and readable by Gemini in BigQuery.

Deputy: All Timesheets report
A new All Timesheets report type is now available in the Deputy source, using the v2 Timesheets API with required start/end date filters. The existing report has been renamed to My Timesheets.

AI agent: token usage progress indicator
The AI agent now shows how much of your token budget has been used in the current session.

AI Integrations page redesigned
The AI Integrations page now displays available tools as visual cards with a 3-step setup guide. Cards show a checkmark if you already have a data flow for that tool.

Dashboard template preview opens in a modal
Clicking a template card no longer navigates away from the list — it now shows a live preview (Power BI) or a full screenshot (other tools) in a modal.

Fixed

Execution error links on Dashboard now clickable
Markdown links in dataflow execution errors were rendering as raw text; they now render as clickable links.

QuickBooks Reports: AI agent edits no longer change columns
The AI agent was incorrectly modifying report source configuration when updating a QuickBooks dataflow. Column selection is now preserved on AI edits.

GoHighLevel: Appointments report now returns complete data
For large locations, individual unresolvable contact IDs were causing entire pages of results to be dropped. The importer now processes contacts one at a time, so partial failures no longer affect the rest of the data.

Confluence source: body columns now return data
The body.storage and body.atlas_doc_format columns were returning empty. The API query now uses the correct parameters, and page sizes were increased.

New

Stripe: OAuth and API key authentication The Stripe source now lets you choose between OAuth and API key when connecting. Both options work in the standard setup flow and via connect-using-link.

Redesigned sign-up page Updated styling aligned to the new design system.

AI-tool-aware sign-up landing page Users arriving from Claude, ChatGPT, or Cursor now see a co-branded page with AI-specific context instead of the generic trial page.

Redesigned OAuth consent screen for AI tool connections The authorization screen now shows a plain-language permission list, workspace selector, co-branded header, and the logged-in email in the footer.

Improved

Looker Studio: one-click template setup Creating a Looker Studio report from a template is now a single step. Clicking the template button creates the report directly — no separate schema load or extra URL parameters needed.

AI agent: dynamic field options now supported The agent can now load and set dynamic dropdowns for sources (e.g. GA4 property lists, account selectors). Previously, sources with dynamic options couldn't be fully configured via the agent.

AI agent: MCP run-dataflow validates before running The run-dataflow tool now checks that a data flow is fully configured before attempting a run, instead of returning success on incomplete setups.

AI agent: MCP get-schema returns column descriptions The get-schema tool now includes column descriptions alongside names and types, giving AI agents more context when working with datasets.

Fixed

Shopify: incremental sync with both date filters Using both created after and created before filters no longer causes a full re-fetch on every run.

Shopify: state saving for nested paginated data Improved checkpoint frequency for orders with line items, products with variants/metafields, and orders with fulfillments — reduces risk of data loss on interrupted runs.

Destinations: connection count shown incorrectly A connection not used in any data flow was shown as active, while the actually-used one showed no flows.

Workflow runs getting stuck in pending state Runs no longer hang when a source or destination changes mid-run.

"Source is not configured" error after fixing an incomplete source Users no longer need to reopen the source step to clear a false error before running.

OAuth redirect broken for new users installing from AI tools New users were sent to the dashboard instead of back to the AI tool after account creation, silently failing the connector install.

"Share credentials" toggle not working for API key auth sources The sharing option now works for all auth types, not only OAuth-based ones.

Run history table: horizontal scrolling on long source names Long source names now truncate with ellipsis; full name shown on hover.

New

AI-assisted data flow setup You can now set up data flows using the AI agent or AI integrations.

New AI destinations: Microsoft Copilot Studio and Gemini Enterprise Deliver your data directly to your enterprise AI tools.

Share a link to connect a source account Workspace admins can now generate a shareable link that lets external users connect their source account without a Coupler.io account. Manage all incoming requests from the Sources page. Learn more

Improved

Klaviyo: Flow performance report split-by options available Daily, weekly, monthly and hourly date aggregations and other split-by breakdowns are available now.

Metabase: updated to API key authentication Metabase now authenticates using an API key. The previous username/password and session token options have been removed. Generate a key in your Metabase account and enter it in the source wizard.

ClickUp: incremental sync for Tasks Tasks now sync incrementally, so only updated records are fetched on each run - reducing sync time and data volume.

LinkedIn Ads: slow Creatives loading The Creatives entity was taking up to 5 minutes to load even for small datasets. Now loads normally.

AI agent: context generation progress indicator A spinner now shows on the context button while context is being generated, so you know it's still processing.

Fixed

Shopify: data preview for Products no longer fails Fixed a bug causing the preview step to fail for the Products entity in the Shopify GraphQL source.

Snowflake: no longer requires CREATE DATABASE privilege for existing databases Snowflake destination now checks whether the target database exists before attempting to create it, so you no longer need the CREATE DATABASE privilege when the database is already there.

Redshift: backslash characters in data no longer cause failures Backslash characters in source data were causing Redshift destination runs to fail. Fixed.

Data sets: hidden columns no longer appear in Join setup Columns marked as hidden were still showing up in column pickers when configuring a Join. They no longer appear.

Integration and destination fixes

• Adobe Commerce (Magento): fixed authorization errors and data retrieval issues affecting some users

• Brevo: row count mismatch between dataset preview and destination output for emailCampaigns resolved

• Monday.com: fixed a 404 error caused by a request format issue

• Outgoing webhooks: fixed a bug where webhooks were not triggering a second data flow after the first run completed

New

Multiple destinations now available You can now send data from a single source to multiple destinations in one data flow. Configure different destinations independently within the same workflow.

AI context management Add and manage context directly in AI tools or Coupler.io Agent to shape how AI understands your data. Provide business definitions, metric explanations, and domain-specific details so the agent delivers more relevant insights. Generate context automatically or paste the one you already store somewhere else.

Supabase source and destination Supabase is now available as both a source and a destination. Connect your Supabase project to import or write data.

Improved

WooCommerce: incremental sync Orders, Products, and Orders with line items now support incremental fetching for faster imports.

ClickUp: incremental fetching for Tasks Tasks now sync incrementally using the date_updated_gt filter.

Google Business Profile: location ID in labels Location labels now include the location ID to distinguish locations with identical names.

TikTok Ads: expanded Audience reports Audience reports now include previously missing metrics and dimensions such as Reach and Province.

PostgreSQL destination: faster large dataset imports Now uses the COPY approach instead of bulk insert for significantly better performance on large data volumes.

Credential removal now disables associated sources When a credential is removed, associated sources are automatically disabled with a notification shown.

Source landing pages: better entity descriptions Source landing pages now display improved descriptions and expanded data entity information.

In-app legal update notifications You now receive in-app notifications when Terms of Use or other legal texts are updated.

Fixed

Google Business Profile: deselecting locations Deselecting one location no longer deselects another for Daily metrics, Monthly used keywords, and Reviews list entities.

Data flows page: filter display bugs Running data flows with a prior successful run now appear only in the Running category. Old data flows no longer incorrectly appear under "No runs."

Workspace: asset reassignment during role changes Users with both org owner and admin roles can now be selected as reassignment targets for data flows and credentials during their creator’s removal.

Search bar: now stays visible when scrolling The search bar on Sources and Destinations pages is now sticky.

Integration fixes Resolved data import issues in Lemlist (Activities pagination), Fillout (Submissions in both sync modes), Cin7 (timeout errors), Float (source not found error), and YouTube (custom report metrics/dimensions not clickable). Fixed CSV destination copy button showing "undefined."

New

Klaviyo: Campaign and Flow performance reports Two new reports are now available in Klaviyo source setup - Campaign performance and Flow performance (totals split by date). Pull campaign and flow-level metrics directly into your data flows.

Customer.io: EU region support You can now select your data center region when connecting Customer.io. If your account is on the EU data center, connections will work out of the box.

MCP: dataset search and embedded skill prompts The AI agent can now search across datasets in larger workspaces to find the right data flow faster. Skills also include embedded prompts for better guidance when invoked via MCP clients.

Improved

Stripe: full refresh sync for all entities Stripe data flows now use a full refresh approach that removes the previous timeout limitations. Large-volume accounts can now reliably sync Charges, Invoices, Balance Transactions, Subscriptions, Customers, Payouts, and more.

GoHighLevel: new filters for Contacts, Appointments, Opportunities, and Tasks Added date range filters and pipeline/status filters across key entities, giving you more control over what data gets exported.

ClickUp: filter Tasks by List and improved Assignee columns You can now filter Tasks by specific Lists. Assignee data is also split into separate name and email columns for easier analysis.

BigQuery: schema autodetect removed for new destinations New BigQuery destinations now use schema transformation by default — the autodetect toggle is no longer shown.

Data flows: filters preserved in URL Active filters on the data flows list now persist in the URL — they survive page reloads and can be shared with teammates.

Destination display improvements on data flow list Destinations on the data flow list now show a cleaner layout with per-destination last run timestamps.

Fixed

Gorgias: Tickets timeout and sync issues resolved Fixed timeout errors and incremental sync problems when importing Gorgias Tickets.

Workspace management: owner/admin deletion bug Fixed an issue where users with owner or admin roles couldn't be selected as reassignment targets during owner deletion.

New

Data flow filtering and explanations Filtering on the data flows list makes it easier to find what you're looking for. Core product sections (data flows, sources, destinations) now include clearer in-UI explanations.

Right sidebar: contextual help for all sources A GitBook-powered help panel is now available in the right sidebar across all sources, providing in-context documentation alongside your current workflow.

Improved

ClickUp: access to archived tasks You can now pull archived tasks from ClickUp alongside active and closed tasks, giving you full access to historical task data.

Lightspeed Retail (X-Series): Sales with line items A new Sales with line items entity makes it easier to link sales data back to product records.

Datadog: RUM Sessions data The Datadog source now supports Real User Monitoring (RUM) Sessions, letting you import session counts, session time, and other RUM event data.

Redesigned sources and destinations experience The Sources empty state now shows categories and ranked lists to help you find the right source faster. The Destination page layout and interactions have been refined for clarity. The setup wizard now includes search for destinations, making it faster to find what you need.

Organization owners page: search and alphabetical ordering The Organization Users page now has a search field, and users are listed alphabetically - easier to manage for large organizations.

Fixed

GoHighLevel: missing custom fields Custom fields in GoHighLevel Contacts and Opportunities no longer go missing intermittently depending on operation order.

Workspace tiles: long name overflow Workspace tiles now correctly truncate long names on wider layouts, preventing overflow and visual glitches.