Which AI models does your code reference, and whose weights are they built on? Paste code or config, or trace a Hugging Face model's declared lineage. Distilled and fine-tuned derivatives count, so deepseek-r1-distill-llama-70b shows up as DeepSeek + Meta. The CLI does the same for a whole repo: npx github:thinx-pro/model-census.
Runs locally in this page. Nothing is uploaded.
Try:
Sends the model ids to the public Hugging Face API (huggingface.co) from your browser. Lineage is what uploaders declared in base_model; the tokenizer check looks for DeepSeek's special tokens, which survive renaming.