Frequently Asked Questions
ScrubMyText text tools, API/MCP usage, ReworkMyText, LockMyAction, CatchMySignal, ApproveMyAction, billing, privacy, and support.
What does ScrubMyText clean?
The main cleaner removes or normalizes common hidden/control characters, non-breaking spaces, whitespace artifacts, and—when requested—selected typographic punctuation. It is designed to preserve wording rather than rewrite prose.
Does text pasted into the browser cleaner leave my device?
The ScrubMyText browser text tools are designed to perform their text operations locally in the browser. Normal website infrastructure, analytics, or advertising services can still receive ordinary request metadata such as IP address and page information.
Does the Document Safety Scanner upload my file?
No. The document scanner is designed to parse the selected file locally in your browser rather than send it to the ScrubMyText API. PDF parsing loads the open-source PDF.js parser code in the browser. The scanner does not OCR image-only content and cannot guarantee that every hidden or malicious instruction will be detected.
Which API tools are free?
The shared free API allowance includes scrub_text, inspect_text, fingerprint_action, check_agent_readiness, list_quality_cards, get_quality_card, and compare_quality_targets. Founding Access unlocks all 49 hosted REST and remote MCP tools for 30 days or 1,000 total calls after one identity verification, with no download, payment card, overage, or automatic subscription. After the grant, the same key remains usable for 500 monthly TrustMyChoice observation and correction writes. ReadyMyAgent checks supplied metadata only, and TrustMyChoice quality cards expose uncertainty rather than certifying a service.
What is ReworkMyText?
ReworkMyText is the deterministic rewriting tool at /rework-my-text/. The browser version runs locally without an LLM; the paid REST and MCP versions expose the same rule engine as rework_text.
What is LockMyAction?
LockMyAction provides stable action identity and persistent idempotency state. An agent can derive and atomically lock a key from structured action arguments, then complete it with the returned lock_id so later retries cannot execute it again.
Does checking an action with LockMyAction reserve it?
No. check_action is read-only. The integration must call lock_action or lock_action_intent and execute only when that call returns safe_to_execute: true.
What is CatchMySignal?
CatchMySignal creates temporary webhook endpoints. An external service can POST an event while the agent is not running, and the agent can retrieve the stored event later through REST or MCP.
How long does CatchMySignal store events?
An inbox lasts 24 hours by default and can be configured up to 7 days. The current default is 25 events per inbox, with a maximum of 100, and each event body is limited to 64 KB of UTF-8 text.
What is ApproveMyAction?
ApproveMyAction creates a secure shareable link for a human to approve or reject a proposed agent action. The agent can poll the approval status before proceeding.
Does ApproveMyAction verify the human's identity?
Not in the current MVP. Possession of the live approval link is treated as authority to decide, so the link should be shared only with the intended reviewer.
Does ApproveMyAction email the reviewer?
Not yet. The current version returns the secure approval URL to the agent/application, which can send it using its existing email, chat, or notification channel.
What counts as one API call?
Each authenticated REST tool request or MCP tools/call invocation counts as one call. prepare_for_agent counts as one call even though it runs multiple internal checks. CatchMySignal also counts each incoming external event delivery as one call. A human approve/reject click does not count.
Are there automatic overage charges?
No. The current paid plans do not automatically bill API-call overages. Access can be limited when the monthly allowance is reached.
What happens to ordinary text sent to the API or MCP server?
Ordinary ScrubMyText text-tool inputs must be transmitted for processing, but the application is designed not to persist those text-tool inputs or outputs in its billing and usage database. The stateful agent-control products intentionally store the limited data required for their function; see the Privacy page.
What is the MCP server?
The remote MCP endpoint is https://api.scrubmytext.com/mcp and the official Registry name is com.scrubmytext/tools. It exposes 49 tools in the current version.
Does ReworkMyText remove AI watermarks?
ReworkMyText substantially changes text when deterministic rules match, but ScrubMyText does not claim to detect, verify, or certify removal of model-level statistical or provenance watermarks. Its percentages measure changes between the input and output, not watermark status.
How do I get help?
Email scrubmytext@gmail.com for billing, subscription, API, MCP, or product support. Do not send passwords, private keys, full API keys, live signal URLs, live approval links, or payment-card numbers.