Extract Evidence
tb_extractExtract claim outcomes from web pages by checking each claim against page content and returning citations with confidence bands. Supports both stateless URL extraction and active browser tab.
Instructions
Extract claim outcomes from a target or opened daemon tab. In a session, omit target to use the active opened tab; claims is always required.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| target | No | The URL or target whose evidence should be extracted. Required for stateless extraction; optional when sessionId points at an opened active tab. | |
| claims | Yes | Required list of claims to verify against the target or active opened tab and support with citations when possible. | |
| browser | No | When true, use the browser-backed acquisition path for this operation. | |
| budget | No | Token budget for snapshot and semantic processing. Default is 512. | |
| mainOnly | No | When true, extract only from the page's main content. | |
| verifierCommand | No | An optional verifier command to run for this operation. | |
| sourceRisk | No | The source risk label to attach to the result or citation. | |
| sourceLabel | No | The human-readable source label to attach to the result or citation. | |
| allowDomains | No | Restrict the operation to this list of allowed domains. | |
| sessionId | No | The browser session to use for this operation. | |
| tabId | No | The specific tab to use inside the session. If omitted, the active tab is used. |
Output Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| open | No | ||
| extract | No | ||
| sessionState | No | ||
| sessionId | No | ||
| tabId | No | ||
| diagnostics | No | ||
| result | No |