# ScrubMyText

ScrubMyText provides deterministic text-integrity tools for humans, software, and AI agents.

Website: https://scrubmytext.com/
Browser tools: https://scrubmytext.com/tools/
AI-agent overview: https://scrubmytext.com/ai-agent-tools/
Developer docs: https://scrubmytext.com/developers/
API: https://api.scrubmytext.com
MCP: https://api.scrubmytext.com/mcp
MCP guide: https://scrubmytext.com/mcp/
MCP registry metadata: https://scrubmytext.com/server.json
OpenAPI: https://scrubmytext.com/openapi.json
Expanded agent index: https://scrubmytext.com/llms-full.txt
Pricing: https://scrubmytext.com/pricing/
API account: https://scrubmytext.com/account/

Tools:
- scrub_text — normalize hidden/control characters and formatting artifacts without rewriting wording. Free + paid.
- inspect_text — exact Unicode, invisible-character, normalization, and mixed-script diagnostics. Free + paid.
- scan_untrusted_text — structural and heuristic risk signals for untrusted text; not a maliciousness verdict. Paid.
- redact_sensitive_text — pattern-based redaction of common credentials and identifiers. Paid.
- compare_text — exact comparison and standard-scrub equivalence. Paid.
- detect_secrets — detect likely credentials without echoing secret values. Paid.
- validate_output — parse JSON and validate common JSON Schema rules. Paid.
- sanitize_agent_context — normalize and optionally redact untrusted context before model ingestion. Paid.
- prepare_for_agent — one-call preprocessing pipeline returning agent-ready text plus a deterministic report. Paid.

Free API access allows scrub_text and inspect_text up to 2,500 calls per public IP per UTC month. Paid plans use Authorization: Bearer smt_live_... and unlock all tools.

ScrubMyText does not claim to remove model-level AI provenance watermarks. Security scans are signals, not safety guarantees.