--- title: "CAT 3.0" description: "CAT release notes 3.0" url: "https://docs.justcat.it/release-notes/recent/3.0/" --- # CAT 3.0 ## New CAT Plans We are introducing new plans for CAT. Our goal is simple: to bring the benefits of automated data testing to a much broader audience. Whether you test data alone or across a whole team, with a budget or without one, there is now a plan for you: **Starter** Starter plan is intended for people who want to **try / evaluate CAT** or want to **use it for free**. In this plan you can use CAT at no cost. There are some limits - you can use only CAT Studio, you can have up to 200 tests for 2 data sources per project etc. (total number of tests is unlimited though, when you split them to more projects). See [Plans comparison](https://docs.justcat.it/compare-plans/ "Compare plans") for details. Nice is that you can use CAT in this plan for any data, in any company, in any environment including Production. **Professional** The Professional plan is intended for people who need to **verify data as part of daily work**. Professional plan removes limits of Starter plan - you can have as many tests and data sources as you like. You can use only CAT Studio. This plan is already powerful enough to have your data completely under control, at friendly price. **Team** The Team plan is intended for team of professionals who need to **collaborate** and **share test results in a professional manner**. It allows teams to stream test results to any output (such as relational database tables) and easily build tailored reporting based on the tests results. **Enterprise** Enterprise plan means the data team can use CAT **without any limits**. Any number of tests and data sources, any outputs. This is the only plan where you can use CAT in **autonomous processes, such as CI/CD**. Check [Plans comparison](https://docs.justcat.it/compare-plans/ "Compare plans") for full details and explanations. If you already use CAT, check [I already use CAT - what changes for me](#i-already-use-cat---what-changes-for-me) section. ## CAT Portal We have a brand new portal for our users: You can buy a CAT plan 💳, manage assignments of seats or provide us feedback 📢. Non-enterprise plans require authentication to the portal. All CAT tools still work locally - that does not change. But without Enterprise key, you need to have an account at CAT Portal and be signed in. There is sign-in dialog in CAT Studio, cmdline tools accept PAT tokens that you can generate in your account at CAT Portal. All tools still only send to the CAT Portal telemetry about feature usage, nothing else. ## AI ### AI agents CAT loves AI agents 😻. AI is everywhere, data testing is no exception. To keep pace with this progress, we are announcing a few important changes: * **Skills** for AI agents: teach your agent how to test data in automated way - all aspects of it. We have a public GitHub repository with the skills: https://github.com/justcatit/cat-skills. Tell about the link to your agent(s) - that should be enough. * To support AI agents, CAT command-line tools now work also on **[Linux and macOS](#linux-and-macos-support)**. * All our **documentation has agents-friendly version** - see [index.md](index.md) for agent friendly version of these release notes. There is Index for LLM/agents. * CAT CLI outputs now have agent-friendly version. ### CAT Pilot You don't use AI agents and still want to leverage AI for creating tests against your data? CAT Pilot in CAT Studio just got a big update: * new models from OpenAI and Anthropic are available * GitHub Copilot support was significantly improved * CAT Pilot's skills are much more powerful * Administrators can now manage CAT Pilot using [group policies](https://docs.justcat.it/how-to-guides/licensing-and-admin/manage-cat-pilot/ "Manage CAT Pilot using group policies"). Whether you use AI interactively or using autonomous agents, testing data now became much easier. ## Linux and macOS support Our command-line tools now have much better support for Linux. Not only CAT CLI, but also CAT PowerShell module and CAT Python module can be used on Linux. All MS Windows dependencies (such as for creating MS Excel report with test results) are gone. You can also run all these tools on macOS - this is a preview feature and we'll appreciate your feedback. ## Native Databricks provider In previous versions of CAT, you had to use ODBC and Simba driver for testing data in Databricks delta tables. Now, support for it comes out-of-the-box. SQL Warehouses, classic compute clusters, serverless / non-serverless - all is supported. Use it from CAT Studio or any CAT command-line tool. ![Databricks support in CAT Studio](databricks-in-cat-studio.png) ## Sort data The `sets match` and `contains` expectation require sorted data. It is still recommended approach to sort data in SQL/DAX queries and provide them sorted for the comparison. Databases perform very well when it comes to sorting and CAT does not need to load data to memory. But in some cases this is not handy: * you have only small datasets to compare (up to ~ 1 million rows) and adding sorting adds no significant performance gain * you struggle with sorting because of different collation in different systems. For such cases, these expectations now support `sort data` (with `sort my data` alias). CAT then loads both datasets to memory and compares it for you. This is explicit opt-in. There is no breaking change, by default CAT still requires sorted dataset, because it is simply better. ![Sort data feature in CAT Studio](sort-data-in-cat-studio.png) ## Documentation Our documentation is now AI friendly, as already mentioned. But we also significantly improved it. * There is a brand new [introduction to CAT](https://docs.justcat.it/what-is-cat/introduction/ "Introduction to CAT") * The [reference section](https://docs.justcat.it/reference/basics/introduction/ "Reference section") was restructured and updated. Every switch/parameter of every tool is properly documented. * How-to guides section is full of tips how to test data. See it on example - [Compare data across systems](https://docs.justcat.it/how-to-guides/test-patterns/compare-data-across-systems/ "Example of a How-to guide"). ## Breaking changes * CAT CLI `run` now exits with a non-zero code (`10`) unless all tests passed. See [Results and exit codes](https://docs.justcat.it/reference/integrations/#results-and-exit-codes "Results and exit codes"). * The name filter is a pattern now: `*` stands for any text, `?` for one character. A filter without a wildcard has to match the whole full name. Where you wrote `-Filter "Airline"` to run every test with *Airline* somewhere in its name, write `-Filter "*Airline*"`. * CAT PowerShell module 3.x requires PowerShell version 7.6 or later (both use .NET 10 under the hood) * Based on users feedback, some MS Excel output columns were renamed and some new were added. * `catcli exec` prints CSV when its output is redirected. ## I already use CAT - what changes for me? If you are on **Enterprise or Business** plan, nothing much changes for you. But we highly recommend you upgrade to 3.0 version: * check [breaking changes](#breaking-changes) * run your tests at least once in a new version and check results * upgrade in a managed manner: * inside automated processes, such as CI/CD, we recommend fixing versions of anything using variables - update your variable to 3.0.0 * if you use both cmdline tools (CAT PowerShell module, CAT Python module, CAT CLI) and CAT Studio, be sure you work with same major.minor version (such as 3.0). If you are on **Personal** plan, your tools will automatically switch to Starter plan - but be aware you can only use CAT Studio now. ----- > **😻:** We wish all **humans** and **agents** a pleasant data-testing experience. Your CAT Team.