In the Export settings, there is now a Failure Condition available to optionally enable: Mark Failed if Zero Rows. This is a useful option if you would like better visibility into your scheduled Exports that may have run with zero rows. This failure would be seen in the Export Logs if the setting is enabled. Otherwise, your Exports with zero rows will be logged as a success.

We’ve upgraded the Shape platform to Facebook Ads API v22!

We’re excited to announce an enhancement to our Facebook Ad Preview experience: the new Force Refresh feature. Users with Manage Templates or Manage Reports permissions can now instantly reload any preview that doesn’t display correctly—whether it’s a “Story unavailable” placeholder, a cookie permission prompt, or a broken'-image icon.

Template Editor & Reports Annotation: Right'-click in the widget and select Refresh Ad Preview Images to reload all visible previews.

Dashboard: Hover over any preview and click the 🔄 icon to reload that single image.

We're pleased to announce the latest version of Pinterest Organic (V5) Provider available in Ninjacat. This iteration includes a couple of improvements:

• When users create their advertiser networks and select their Pinterest network, all boards will be imported. The user can then filter the boards at the widget level.
• To help visualize data within the template editor when there are too many data requests, which limits the amount of data, specifically when trying to bring in Pin reports, we've added a toggle of "Limit Data Requests".

In addition to these updates, we've added the following available fields:

Dimensions:
Pin Title and Pin Description

Metrics:
Pin Engagement Rate (lifetime), Pin Engagements (lifetime), Board Followers (lifetime), Pin Impressions (lifetime), Pin Clicks (lifetime), Pin Outbound Clicks (lifetime), and Pin Saves (lifetime).

Our PDF reports were intermittently missing Facebook ad images due to too many simultaneous image-generation requests, which slowed processing and stalled previews.

We have made significant updates to our Facebook Ad Preview Service, resulting in increased stability and a smoother user experience. Our new system pre'-generates images using advanced caching mechanisms before a report is created, ensuring that images appear reliably. Although this update required a substantial redesign of our backend processes, it now delivers more consistent image previews.

We have added a Custom Date Range option for one-time Re-sync triggered exports (navigate to an existing Export --> Resync --> Export Custom Date Range). This is useful if a user wants to export data from a specific time frame for audit purposes or if a historical export had run with incomplete data so a resync of certain dates is required.

Some date range rules in place:
-Start Date can be any time between now (today) and 2 years ago.
-End Date can be any time between now (today) and 2 years ago.
-Start Date must be on or before End Date
-End Date must be on or after Start Date

We have given all users access to our SQL Transform Builder in Data Cloud. SQL Transforms allow users to manipulate and combine Datasets within NinjaCat using SQL as well as our AI-powered chat. This feature enables users to apply complex transformations, joins, unions and aggregations beyond basic filtering and roll-ups. Learn more about how to use our SQL Transform Builder here: https://docs.ninjacat.io/docs/transform-data-with-sql

For anyone who had utilized our "Unions" combine option, we have retired this feature as it wasn't quite meeting the needs, whereas our SQL Transform Builder allows for much easier handling of complex transformations. Any existing "Union" datasets that had been creating through the Unions feature will still function appropriately.

You are now able to generate PDFs from a convo w/ an Agent, simply by asking the Agent to generate one for you. This is helpful if you're looking for a way to export the Agent's analysis into a shareable format either with teammembers or with a customer.

Requirements for generating a PDF in-convo:
-Code Interpreter must be enabled for that Agent
-Not required, but does seem to work best with the following models so far: Claude 3.7 Sonnet and GPT 4.1.

Over the past week, we have added support for OpenAI's newly released models, which joined our existing options of GPT-4o, o3-mini and Claude 3.7 Sonnet:

-GPT 4.1 - OpenAI's Flagship GPT Model for complex tasks (an upgrade to GPT-4o)
-GPT 4.1-mini - Balanced for intelligence, speed and cost
-GPT 4.1-nano - Fastest, most cost-effective GPT 4.1 model
-o4-mini (w/ Low, Medium and High reasoning levels) - smaller model optimized for fast, cost-efficient reasoning
-o3 (w/ Low, Medium and High reasoning levels) - most powerful reasoning model

With this additional support, we have also revamped our model selection dropdown menu in the Agent Builder, giving you a cleaner, more helpful menu of the NinjaCat-supported model options. With this, we have also made the default model for new Agents be Claude 3.7 Sonnet as we've seen this model work well with many use cases. However it is still up to the user to test out other models with their Agent configuration and select the best for that Agent's task. Learn more about these models through OpenAI's help documentation here:
-Introducing 4.1 Models: https://openai.com/index/gpt-4-1/
-Introducing o3 & o4-mini Models: https://openai.com/index/introducing-o3-and-o4-mini/

We are pleased to announce that we have successfully upgraded our Criteo API to the latest stable version of 2025.01. This update ensures our system remains current and efficient, providing you with the best possible experience.