Compute pairwise text similarity, find near-duplicates, and cluster a batch of texts (2–300 items). Character n-gram cosine similarity: language-agnostic, robust to small edits, no model calls. Returns similar pairs above a threshold, connected-component clusters, and the full matrix for small batches. POST a JSON body like: {"texts":["hello world","hello world!","unrelated"],"threshold":0.8}