TurnitHuman

Detect Ai Content

Detect AI-generated content in text sentences. This tool wraps the existing /gundua API endpoint. Args: sentences: List of sentences to analyze (1-500 sentences) api_key: Optional API key for authentication Returns: JSON string with detection results from the backend API

Start Here Workflow

START HERE: Returns humanization workflow instructions. CRITICAL: Use this before calling humanize or batch_humanize. Read all instructions completely — do not skip any part. Args: task_description: What you want to humanize (e.g., "full document essay", "screenshot highlighted text", "newly added sections") Returns: Plain-language workflow instructions including preservation rules, chunking strategy, and quality control steps.

Humanize

Humanize AI-generated text to make it sound more natural and human-like. This tool transforms text through advanced AI processing to remove AI patterns and make content sound authentically human while preserving meaning. Args: text: Text to humanize (any length - Claude chunks intelligently) style: REQUIRED - Writing style: casual, conversational, highschool, university, professional, phd Returns: JSON string with humanized text and word counts

Batch Humanize

Humanize multiple text chunks OR get multiple style variants of same text in parallel. TWO USE CASES: 1. Multiple chunks, one style: chunks=["chunk1", "chunk2"], style="university" 2. One chunk, multiple styles: chunks=["same text"], styles=["university", "highschool", "professional"] - Use this for quality control - get multiple style variants to mix best sentences - COST: Credits = words × number_of_styles (200 words × 3 styles = 600 credits) Args: chunk

Echo Test

Simple echo test to verify MCP connection. Args: text: Text to echo back Returns: Echo response