KiHustle Jaccard Similarity Token

Choose this endpoint when you need a lightweight, fast, token-based similarity metric between exactly two text strings. It is ideal for deduplication, near-duplicate detection, or keyword overlap checks where you want a simple set-intersection metric rather than semantic embedding similarity. Prefer it over embedding-based similarity when interpretability and speed matter more than capturing semantic meaning.

• Missing text_a or text_b fields returns an error
• Empty string inputs may return a score of 0 or an error
• Very long texts may cause timeouts or unexpected results
• Non-string inputs may cause processing failures
• Payment not included or insufficient USDC may result in 402 rejection