Status and result tools

MCP action tools that create persisted runs return a run_id.

Follow the returned next_action when it is present. It points to the correct tool for the run shape.

Status Tools

Tool	Use it for
get_sleepwalker_visibility_run_status	Poll direct AI Visibility runs.
list_sleepwalker_visibility_runs	List direct AI Visibility runs.
get_sleepwalker_content_run_status	Poll direct Content Intelligence runs.
list_sleepwalker_content_runs	List direct Content Intelligence runs.
get_sleepwalker_run_result	Fetch saved Console run details, including saved-test Content Intelligence reruns.
get_sleepwalker_prompt_response	Fetch one AI Visibility prompt response chunk.

Required scope:

runs:read

Polling Pattern

create run
  -> store run_id
  -> poll status every 10 to 30 seconds
  -> stop when terminal
  -> fetch heavier result detail only if needed

MCP create and status responses can include next_action. Treat it as the recommended next tool call for the current run state.

next_action.version is currently 1. Supported next_action.type values are:

Type	Meaning
poll_status	Poll the named status tool after the suggested delay.
fetch_result	Fetch the full result only if the user needs detailed evidence.
explain_terminal_status	Explain a failed or cancelled terminal state before retrying.

Common Statuses

Status	Meaning
queued	Accepted and waiting to start.
running	Work has started.
completed	Work finished and result data is available.
failed	Work could not complete.
cancelled	Work was not run or was stopped.
partially_queued	Some requested probes were queued and some were skipped.

See Enums for shared status values.

MCP accepts common status wording such as done, complete, and in progress, then normalizes it to the canonical status.

Include Flags

Status and result tools return compact data by default.

Tool	Flag	Use when
get_sleepwalker_visibility_run_status	include_probes	You need per-probe status rows.
get_sleepwalker_visibility_run_status	include_results	You need available probe result data.
get_sleepwalker_content_run_status	include_result	You need the full Content Intelligence result.

Use include flags deliberately. Large payloads are harder for MCP clients to handle.

Prompt Response Drilldown

For any AI Visibility run (saved-test or direct-URL), fetch individual prompt responses with:

get_sleepwalker_prompt_response

This is the preferred way to read a large run: pull one answer at a time instead of the full include_results payload. The response includes the model that answered, the full (capped) text, and citations.

Inputs:

Input	Type	Notes
run_id	string	Required run ID.
result_index	integer	Fetch by flattened result index (same order as the status ca_result).
platform	string	Fetch by platform. Canonical values and common labels such as ChatGPT or Grok are accepted.
prompt_index	integer	Fetch by prompt position.
max_chars	integer	1 to 25,000.
offset	integer	Continue long text from an offset.

Example Prompts

Poll this Sleepwalker run until it completes, then summarize the result.

Fetch the first prompt response with citations for this visibility run.