Limits

Sleepwalker is designed for agent and API reliability.

Payload Budget

Some clients have practical response-size limits. Sleepwalker returns compact payloads by default.

Large artifacts are exposed through summaries and drilldowns:

• Run summaries return compact counts and IDs.
• Prompt responses are fetched one at a time.
• Page serialization returns bounded chunks.
• Status endpoints avoid embedding oversized result payloads.

Batch Runs

Run bulk AI Visibility work as one queued batch action, not hundreds of separate tool calls.

Good:

Create one run with 100 prompts x 4 platforms.

Bad:

Call the tool 400 times from an agent loop.

Partial Execution

If a user requests more probes than their current credits can fund, Sleepwalker can still run the funded probes and mark the rest as not run because of insufficient credits.

The response shows what ran, what did not run, and why.

API Limits

Area	Limit
Visibility prompts per matrix run	100
Visibility platforms per matrix run	4
Visibility probes per run	400
Visibility prompt length	500 characters
Visibility competitors	10
List endpoint page size	100
Page serialization max_chars	25,000

Rate Limits

No public fixed request-per-minute limit is currently exposed in the API.

Sleepwalker may still reject abusive or unsafe traffic.