📚 Editorial Agent

The Editorial Agent is a local-first AI assistant that watches a folder, reads incoming .pdf and .docx documents, intelligently chunks them, reviews each part using GPT-4o, and generates a detailed editorial feedback report — all fully automated.

✨ Why Use This?

• Designed for teachers, authors, curriculum designers, and editors
• Helps polish drafts, books, lesson plans, or study material
• Offers structured, human-grade editorial feedback
• No manual uploading — just drop a file and let the agent do the work

🛠 How It Works

input_documents/  →  🔍 chunked and reviewed →  output_reports/
                                   ↓
                         logs/editor_logbook.csv

Main Components:

• editor_agent.py – Orchestrates the full flow
• editor_chunker.py – Breaks long documents into smart chunks
• editor_reviewer.py – Sends each chunk to OpenAI with a powerful editorial prompt
• editor_report_writer.py – Compiles a polished Word doc with feedback

📦 Getting Started

1. Clone the Repo

git clone https://github.com/yourusername/editorial-agent.git
cd editorial-agent

2. Install Dependencies

pip install -r requirements.txt

3. Add Your API Key

Create a .env file:

OPENAI_API_KEY=sk-...your-key-here...

4. Run the Agent

python scripts/editor_agent.py

Drop files into input_documents/ and watch them get processed. Reports are saved to output_reports/.

📄 Sample Output

See sample_output/report_example.docx

🔍 Example Use Cases

• Reviewing K-12 textbooks
• Improving instructional design for teachers
• Editing fiction or non-fiction drafts
• Flagging clarity or grammar issues in training content

💡 Roadmap Ideas

• Export to PDF
• Live web interface
• Team dashboard for tracked feedback
• Built-in style guides

📜 License

MIT — use it, build on it, share it.

🧠 Inspired by a Live Agent Build

This project was built step-by-step using AI. See agent_building_guide.md for how to create your own AI agent from scratch.

---

Let the agent edit while you create.