Skip to content

Content Intelligence

Content Intelligence tools help agents evaluate a page against market context and return recommendations.

Use one-off tools for immediate answers. Use persisted runs when the result should be saved and pollable.

ToolPurposeScope
discover_sleepwalker_content_trendsFind relevant content trends for a URL.content_intelligence:trends:discover
score_sleepwalker_contentScore a URL without saving a run.content_intelligence:content:score
create_sleepwalker_content_runQueue a persisted Content Intelligence run.content_intelligence:runs:create
list_sleepwalker_content_runsList direct Content Intelligence runs.runs:read
get_sleepwalker_content_run_statusPoll one Content Intelligence run.runs:read

Use this when the user wants market context only.

Inputs:

InputTypeRequiredNotes
urlstringYesPage URL.
extraction_modestringNoOptional extraction profile. Omit unless instructed otherwise.
languagestringNoDefaults to en.
countrystringNoDefaults to US.

Output includes compact industry, trend, and citation data.

Credits: 1 when completed successfully.

Use this when the user wants a score, gaps, and recommendations without saving a run.

Inputs:

InputTypeRequiredNotes
urlstringYesPage URL.
extraction_modestringNoOptional extraction profile. Omit unless instructed otherwise.
industrystringNoOptional industry label.
trendsobject[]NoOptional trend set from trend discovery.
languagestringNoDefaults to en.
countrystringNoDefaults to US.

If the agent already called discover_sleepwalker_content_trends, it should pass the returned compact trends into score_sleepwalker_content.

Credits:

ModeCredits
With supplied trends2
Without supplied trends3

Use this when the result should be persisted and retrievable later.

Inputs:

InputTypeRequiredNotes
urlstringConditionalNew direct-URL run. Send exactly one of url or test_id.
test_idstringConditionalRerun an existing saved Content Intelligence test.
analysis_depthstringNoscore or full. Defaults to full.
languagestringNoDefaults to en.
countrystringNoDefaults to US.
idempotency_keystringNoRetry-safe key.

Use url when the user wants to analyze a new page.

Use test_id only when the user explicitly asks to rerun an existing saved Content Intelligence test.

The create response includes a next_action field when the run is queued or running. Agents should follow that exact tool and arguments instead of creating a duplicate run. Direct URL runs normally poll get_sleepwalker_content_run_status; saved-test reruns may poll get_sleepwalker_run_result because they reuse the Console result path.

DepthUse whenCredits
scoreThe user needs scoring and recommendations.3
fullThe user needs the complete Content Intelligence run output.6

Use:

get_sleepwalker_content_run_status

for direct URL runs. For saved-test reruns, follow the next_action returned by create_sleepwalker_content_run; it can point to get_sleepwalker_run_result.

Inputs:

InputTypeRequiredNotes
run_idstringYesRun returned by create.
include_resultbooleanNoDefaults to false.

Poll every 10 to 30 seconds.

When a compact status response is complete, next_action points back to get_sleepwalker_content_run_status with include_result: true so the client can fetch the full Content Intelligence output only when needed.

Score https://www.example.com and show the top recommendations.
Discover trends for this page first, then score the page against those trends.
Create a full Content Intelligence run for this URL and poll until the result is ready.