Waka - Detecção de Texto Gerado por IA
thehackersplaybook/waka
Detecção confiável de texto gerado por IA com explicabilidade
Ver no GitHubAPI service que detecta texto gerado por modelos de IA usando GPT e Claude, fornecendo scores de probabilidade e explicações em linguagem natural. Otimizado para baixa latência e suporta múltiplos idiomas, sendo ideal para moderação de conteúdo, verificação acadêmica e análise de autenticidade em tempo real.
Instalação
npx convez-ai thehackersplaybook/waka1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Moderação de conteúdo em redes sociais e fóruns
- Verificação de integridade acadêmica em trabalhos e pesquisas
- Fact-checking e validação de autenticidade de artigos
- Pesquisa e benchmarking de modelos de IA generativa
Veredito

Ferramenta experimental mas promissora para detecção de conteúdo sintético. O uso de LLMs para explicabilidade é inteligente, porém a precisão depende fortemente do modelo subjacente escolhido — adequada para prototipagem e casos não-críticos, não recomendada ainda para produção em missões críticas.
README
A tradução para o português deste README ainda está na fila.
Waka: Detect AI-Generated Text 🐲
Disclaimer 🚦: This project is currently in an experimental stage, use at your own discretion. Please do not use Waka for mission-critical applications yet.
Overview
Waka is an API service designed to detect AI-generated text using Gen AI models by Open AI and Anthropic. Whether you're verifying content authenticity, combating misinformation, or simply curious about the origins of text, Waka provides a reliable and efficient solution for distinguishing between human-written and AI-generated content.
Features
- ✨ Scoring & Reasoning: Waka generates a 'ai generated score' and a natural language reason to justify the score. This design decision was taken in favour of our XAI (Explainable AI) initiatives.
- ⏳ Fast Responses: Optimized for low-latency detection, making it suitable for real-time applications. The bottleneck is the LLM, so you can switch to more lightweight models like GPT 4o Mini for quicker responses.
- 🛠️ Scalable: Supports high-volume requests, perfect for enterprise-level integrations. Since you use your own key, this depends on your LLM Provider Plan.
- ⚙️ Customizable Thresholds: Tailor detection sensitivity to your specific use case. (WIP)
- 🌍 Multi-Language Support: Works with text in multiple languages to ensure wide applicability. Most LLMs include training data written in languages other than English, so this feature comes out of the box.
- 🔧 Developer-Friendly: Intuitive API with comprehensive documentation and easy-to-use endpoints. There's a Postman Collection provided below. In future, we plan to add a detailed documentation and setup page to make life easier for adopters.
Use Cases
- 🕵️♂️ Content Moderation: Identify AI-generated spam or synthetic content in forums, social media platforms, or comment sections.
- 🎓 Academic Integrity: Detect AI-written assignments or research papers.
- 🔧 Fact-Checking: Verify the authenticity of news articles or user-generated content.
- 🤖 AI Research: Analyze and benchmark the performance of AI-generated text models.
- 🔎 Personal Use: Satisfy curiosity about the authenticity of a piece of text.
Setup Instructions
Prerequisites
Ensure you have the following installed on your system:
- 💻 Node.js: Download and install Node.js
- 🔧 npm: Comes bundled with Node.js
Installation
- Clone the repository:
git clone https://github.com/ythehackersplaybook/waka.git cd waka - Install dependencies:
npm install
Setup Postman Collection
Download the Postman Collection (thp-waka.collection.json) and play with the APIs yourself. (TODO: add this)
Running the Service
- Set up environment variables:
- Create a
.envfile in the root directory. - Add the following keys:
OPENAI_API_KEY=your-openai-key ANTHROPIC_API_KEY=your-anthropic-key
- Create a
- Start the development server on
http://localhost:3000:npm start - Build the project:
npm run build - Run tests:
npm run test
Roadmap
- 🌍 Expand multi-language support to cover all major global languages.
- ⚙️ Add batch processing endpoints for analyzing large datasets.
- 🕌 Implement a browser-based dashboard for non-developers.
- 🔍 Introduce AI attribution for identifying the specific AI model used to generate text.
Contributing
We welcome contributions! Please follow these steps:
- 🔄 Fork the repository.
- 🕳️ Create a new branch for your feature or bug fix.
- 🔗 Submit a pull request with a detailed description of your changes.
Support
For questions or support, reach out to us at ✉️ thehackersplaybook0@gmail.com or visit our documentation site.
License
🔒 Waka is licensed under the MIT License. See LICENSE for more information.
Unmask the truth, one text at a time with Waka! 🐲