All analysis with Allye is performed entirely in your local environment. Your raw data is never shared with large language models (LLMs) or uploaded to the cloud, ensuring that even highly sensitive data can be analyzed securely.
You do have the option to share header information—such as column names—with the LLM. This helps the model generate Python code more effectively.
However, when using the AI Reporting feature to generate textual summaries of your statistical analysis, certain results—such as visualized charts and statistical metrics (e.g., p-values)—will be shared with the LLM. These are not stored permanently and are deleted immediately after use.

If an error occurs while building a workflow in Allye, a confirmation dialog will appear asking whether you’d like to send the error log to us. The log is only shared with us if you give explicit permission, and it will be used solely to help improve the product. Your raw data and any Python code used for analysis remain strictly within your local environment and are never stored elsewhere. In addition, we store and manage user account registration information.

Not at all. Allye is designed to be fully usable without any coding skills. You can build workflows using a no-code interface. However, for advanced users, there is also the flexibility to write or edit Python code as needed.

Most of Allye’s core features run entirely offline in your local environment, so no internet connection is required. However, features that rely on AI models—such as code generation or automated report writing—may require a temporary internet connection when used.

Yes. You can save your workflows locally and reopen them anytime. Allye also allows you to reuse and share workflows easily, making it convenient to replicate or collaborate on analysis.