> ## Documentation Index
> Fetch the complete documentation index at: https://docs.levanto.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Grounding

> Optional web search before deciding on low-confidence requests.

Attach a `grounding` block to a Yes/No, Choice, Scale, or Tags request to let Sage fetch fresh web evidence before deciding. Useful for recency- or fact-heavy questions.

## Request

```json theme={null}
{
  "content": "Has there been a major Cloudflare incident this month?",
  "question": {"id": "incident", "kind": "yesno", "instructions": "Is this likely true?"},
  "grounding": {"trigger": "low_confidence", "confidence_floor": 0.8}
}
```

## Options

| Field              | Description                                                                         |
| ------------------ | ----------------------------------------------------------------------------------- |
| `trigger`          | `never`, `low_confidence` (default), or `always`                                    |
| `confidence_floor` | For yes/no, default `0.80` requires roughly `p ≤ 0.10` or `p ≥ 0.90` to skip search |

## Response

The response gains `grounding_meta` with `triggered`, `queries`, `sources`, and timing.

<Warning>
  Grounding requires the server to be configured for it. If it is not, omit the block.
</Warning>

For Tags, `low_confidence` triggers when **any** tag's `confidence` is below the floor (weakest-link); Sage then re-scores all tags on the expanded content.

## Pricing

When Sage actually runs a search (`grounding_meta.triggered` is `true`), that decision costs **an extra cent** on top of the usual per-decision price. No search, no surcharge. See [Pricing](/pricing) for the full breakdown.

## Batch placement

On `/decide`, `grounding` is a top-level field. On [/decide/batch](/decision-model/batch), there is no group-level grounding — each question in a group's `questions` list carries its own optional `grounding` block (`sort` questions still reject it).