Tried the **semantic-find cookbook** on supplier replies for Factory Twelve, an apparel sourcing platform, using `jev-1.13.0`.

Tried the **semantic-find cookbook** on supplier replies for Factory Twelve, an apparel sourcing platform, using `jev-1.13.0`. Three questions over the same state in one System One request: - **Choice — `where`:** select the most relevant clause ID. - **Noul — `exists`:** probability the reply supplies at least one requested detail. - **Choice — `coverage`:** classify the answer as `complete`, `partial`, or `absent`. Existence is separate because clause Choice always produces a winner. Coverage is separate because Noul near 0.5 means uncertainty—not a partially complete answer. Code preserves source IDs, versions, hashes and character offsets, then copies the selected text verbatim. No generated quotations. **First run: 12 synthetic development cases** - Existence: **12/12** correct at a 0.5 threshold. - Clause selection: **8/8** on answer-bearing cases. - Highlights suppressed for all **4 absent cases**. - Coverage: **9/12** correct. - Mean latency: **455 ms/request**. - Usage: **9,742 input / 1,548 output tokens**. All three coverage errors selected the correct clause: - `$