🤖 Multi-Agent AI Blog Generation Crew

Welcome to the BlogGenerationCrew project. This is a production-grade, multi-agent AI system engineered using the CrewAI framework. It leverages a hierarchical management structure to orchestrate autonomous agents that research, draft, and review long-form technical content.

Unlike standard templates, this repository has been optimized to handle complex, high-density reasoning loops entirely through free-tier NVIDIA NIM infrastructure, bypassing common token-ceiling and gateway connection limits.

---

🧠 System Architecture & Workflow

This project implements a Hierarchical Process (Process.hierarchical) to mimic a real-world corporate data operations team:

1. Team Leader (Crew Manager): Ingests the global assignment topic, dynamically establishes a project roadmap, structures tasks, and handles serialization loops.
2. Expert Researcher: Consumes a custom RAG-driven SerperDevTool to search the live web, parsing snippets and chunking complex datasets before passing data back.
3. Expert Blog Writer: Transforms raw research packets into an expansive, detailed, long-form draft.
4. Expert Blog Reviewer: Acts as a quality assurance gate, evaluating structure, tone, and analogies, providing structural feedback scores before finalizing production.

---

⚡ Features & Engineering Optimizations

• NVIDIA NIM Integration: Configured natively to route high-density operations through meta/llama-3.1-70b-instruct.
• Self-Healing Parsing: Integrated resilient schema error catching to gracefully recover from transient multi-agent JSON parsing anomalies (Action Input key-value slips).
• Gateway Timeout Buffers: Tailored underlying HTTP transport socket boundaries (connect_timeout and read_timeout mapped to 600s) to fully mitigate cloud 504 Gateway Timeout loops during deep context processing.
• Extended Token Runways: Pushed generative runtime horizons (max_tokens=4000) to unlock comprehensive, 1,500+ word output generation.

---

🛠️ Installation & Setup

1. Prerequisites

Ensure you have Python 3.10 up to 3.12 installed. This project uses uv for lightning-fast, isolated dependency management.

2. Clone and Install Dependencies

# Clone the repository
git clone [https://github.com/ArindamDeka09/blog-generation-crew.git](https://github.com/ArindamDeka09/blog-generation-crew.git)
cd blog-generation-crew

# Install litellm and project dependencies using uv
uv sync