---
title: "tool si.ask.semantic_search"
description: "a text or a subject's vector, the model, a floor and a population: the units above the floor."
section: reference
address: /reference/tools/si-ask-semantic_search
contract: 0476e35e6e275db5
---

# tool si.ask.semantic_search

a text or a subject's vector, the model, a floor and a population: the units above the floor.

[figure: Over MCP, si.ask.semantic_search takes SemanticSearchInput and returns a card whose body is SemanticSearchAnswer.]

the units above the floor. With them: the model, the floor, the measured recall of quantized embeddings, and the tier. The call needs the ask grant at the strategy rung.



| | |
|---|---|
| input | [SemanticSearchInput](/concepts/semanticsearchinput) |
| result | a card whose body is [SemanticSearchAnswer](/concepts/semanticsearchanswer) |
| read-only | yes |
| destructive | no |
| card address | `ui://scale-intelligence/cards/SemanticSearchAnswer` |
| HTTP route | [`/v1/ask/semantic_search`](/reference/ask/semantic_search) |





| field | type | required | note |
|---|---|---|---|
| `text` | `text` | no |  |
| `subject` | [UnitRef](/concepts/unitref) | no |  |
| `model` | [Model](/concepts/model) | yes |  |
| `floor` | `share` | yes |  |
| `population` | [Predicate](/concepts/predicate) | yes |  |