alumnium-hq/alumnium

AI Summary: This issue requests the ability to configure a custom base URL for Ollama, allowing it to connect to remote deployments instead of being limited to local instances. The proposed solution involves using an environment variable, `ALUMNIUM_OLLAMA_BASE_URL`, to specify the remote endpoint, similar to how other cloud service integrations handle base URLs.

AI Summary: Modify the Alumnium AI-powered test automation library to capture screenshots of specific areas (defined by selectors) instead of full-page screenshots when using `area.get` or `area.check` with `vision=True`. This improvement aims to enhance stability and reduce data sent to LLMs.

AI Summary: Improve Alumnium's Appium locator strategy to handle non-unique locators. This involves generating XPaths, checking against the accessibility tree, and handling multiple matches by using element IDs and appending indices to XPaths. The goal is to make the `swag_labs_test.py` example pass on Appium.

AI Summary: Integrate the ability to use Mistral large language models (LLMs), specifically the Mistral-small 3.1 24B model accessed via an API, within the Alumnium AI-powered test automation framework. This requires adding API interaction capabilities to Alumnium to handle requests to the Mistral API and integrating the LLM responses into the existing test automation workflow.