tutorials
Things you can actually run
Every tutorial here declares what it depends on, which version its steps were run against, and the date they were last run. When that date goes stale the page says so before the first step — silent staleness is a false claim on a published page.
Run a sentence-embedding model in the browser, with no server and no API key
verified · Transformers.js 4.2.0, ONNX Runtime Web 1.26.0-dev.20260416-b7804b056c, all-MiniLM-L6-v2 (ONNX) 2025-07-22
Two mirrors of Llama 3.1 ship identical weights and different prompts, and one tells the model it is July 2024
verified · Hugging Face Hub tokenizer_config.json for NousResearch/, unsloth/ and meta-llama/ Llama-3.1-8B-Instruct as served 2026-08-28, Transformers.js 4.2.0, llama.cpp GGUF version 3; tokenizer.chat_template in bartowski/Meta-Llama-3.1-8B-Instruct-GGUF
Read what is actually inside a 4.9 GB model file by downloading 0.16% of it
verified · Hugging Face Hub huggingface.co as served 2026-08-28; Accept-Ranges: bytes on resolve/ URLs, llama.cpp GGUF version 3, general.quantization_version 2, Quantization bartowski/Meta-Llama-3.1-8B-Instruct-GGUF, repo sha bf5b95e96dac0462e2a09145ec66cae9a3f12067, lastModified 2024-12-01
Watch the OpenRouter model catalog change, with no key and no dependencies
verified · OpenRouter /api/v1/models as served 2026-08-28, total_count 398