Module overview
An AI that stays on your machine, useful from the first question
Ambre AI is the artificial intelligence engine built into PANACHES: conversational chat with memory and searchable history, code and documentation generation from the editor (function, class, script, API, test, refactor...), a local knowledge base (RAG) to query your own documents with citations and a confidence score, seven AI personality profiles and a reusable prompt library. Models run on your machine, with automatic GPU detection, no Internet connection required.
What the module brings
- Chat with AI, with history, search and conversation export.
- Generate code, tests, documentation and technical explanations from the editor, choosing the type and style.
- Query your own locally indexed documents (RAG knowledge base), with citations and a confidence score, without sending that content to a forced cloud.
- Choose among seven AI personality profiles and save your favorite prompts in a reusable library.
- Switch local models on the fly (Gemma, Mistral, LLaMA, Qwen...) depending on the task and your machine's power.
- Browse, search and filter the model catalog, save favorites or manually select a compatible model file.
- Download a model voluntarily with progress tracking and format validation: no heavy model is installed without your consent.
- Tune threads, batch size and GPU layers for your hardware, with GPU/CPU detection and asynchronous loading.