Show loading state on Generate button while recipe runs.

Add dual recipe backends with OpenAI as the default.

Stop tracking plans and Cursor config in git.

Update recipe prompt example and steps formatting rule.

Relax pantry validation and clarify empty extras in the prompt.

Add a taste-focused line to the recipe system prompt.

Revert "Show Epicure suggestions while the recipe is generating."

Show Epicure suggestions while the recipe is generating.

Strip Qwen special tokens before parsing recipe JSON.

Switch to Qwen2.5-3B-Instruct with CPU inference support.

Switch recipe generation to Qwen2.5-7B-Instruct.

Remove explicit HF token from public model loading.

Fix ingredients display and pin kernels for gpt-oss loading.

Load gpt-oss-20b directly from the Hugging Face Hub.

Switch recipe generation to local gpt-oss via Transformers.

Use full Epicure Core vocabulary in ingredient dropdown.

Add Epicure paper and model citations to the UI and README.

Simplify recipe system prompt rules.

Switch recipe generation to OpenAI gpt-4o-mini.

Log full LLM response on invalid JSON.

Relax recipe ingredient rules and add failure logging.

Switch default recipe model to Arch-Router-1.5B.

Add HF Inference recipe generation and wire it to Generate.

Integrate Epicure Core nearest-neighbor shopping list suggestions.

Ignore epicure-explorer nested repo.

Add ingredient suggestion logic and wire it to Generate.

Add ingredient multi-select form with generate button.

Test commit

initial commit