lackey43/humanizer

TEXT GENERATIONPricing:Input $1.2 / Cached $0.24 / Output $4.8Concurrent Unit Cost:1Model Size:12BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Oct 15, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

lackey43/humanizer is a 12 billion parameter model fine-tuned to rewrite AI-generated text into more human-like prose, while preserving all factual details. Developed by lackey43, it supports both English and Chinese, and is optimized for local execution with a 32768 token context length. Its primary use case is to enhance the naturalness of AI-written drafts such as emails, essays, and reports, making them indistinguishable from human-authored content.

Loading preview...

humanizer: Making AI Text Sound Human

humanizer is a 12 billion parameter model designed to transform AI-generated drafts into natural, human-sounding text. It meticulously preserves all factual information, including numbers, units, dates, names, and quotes, without adding new content. This model is particularly effective for emails, essays, reports, and forum posts in both English and Chinese, and is optimized for local, offline execution.

Key Capabilities

  • Human-like Rewriting: Rewrites AI-generated text to sound more natural and less robotic, varying sentence length and cutting hedging.
  • Fact Fidelity: Ensures all critical factual details (numbers, dates, names, quotes) remain unchanged, with 95% of English rewrites having no factual problems in evaluations.
  • AI Detector Evasion: Achieves a 95% "judged human" rate by Originality.ai at its strictest setting, significantly outperforming previous versions and other baselines.
  • Multilingual Support: Proficient in rewriting both English and Chinese texts.
  • Optimized Quantization: Features custom quantization-aware trained GGUF files (2-bit, Q3, Q4_K_M) that offer superior quality at smaller file sizes compared to standard llama.cpp quantizations, enabling efficient local deployment on devices with as little as 8GB RAM.
  • Local Execution: Designed to run efficiently on local machines, with dedicated apps for Mac (Apple silicon) and Windows, and support for llama.cpp, MLX, transformers, and vLLM.

Good For

  • Content Creators: Enhancing the naturalness and readability of AI-assisted writing.
  • Professionals: Humanizing AI-drafted communications like emails, reports, and presentations.
  • Developers: Integrating into applications that require human-quality text output from AI models, especially for local deployment.
  • Users with Privacy Concerns: Performing text rewriting entirely offline without sending data to external services.
  • Researchers: As a base for further fine-tuning on specific humanization tasks, leveraging its Apache 2.0 license.

Finetunes
1 models