Agent Core v1.0.0

🚀 Initial Release: Agent Core v1.0.0

Features

• Persistent Memory: Redis-backed conversation history with sliding window
• Self-Learning RAG: Automatic research ingestion from ArXiv, Google Scholar, and web
• Dynamic Prompts: Self-evolving system prompts with versioning
• Multi-Model Support: Compatible with OpenRouter LLMs (default: Mistral Medium 3.1)
• RunPod Ready: Full serverless deployment configuration included

Components

• TapeAgent-inspired memory architecture
• Multithreaded PDF processing pipeline
• BM25 + FAISS hybrid search
• RQ-based background task processing
• Session-based conversation management

Deployment Options

• Local: Run with Docker Compose
• RunPod Serverless: Deploy with included hub.json configuration
• Docker: Standalone container with bundled Redis

Requirements

• Python 3.11+
• Redis 6.0+
• OpenRouter API key

Quick Start

```bash

Clone repository

git clone https://github.com/phxdev1/agent-core.git
cd agent-core

Set up environment

cp .env.example .env

Add your OPENROUTER_API_KEY to .env

Run with Docker

docker-compose up

Or run locally

pip install -r requirements.txt
python agent_chat.py
```

RunPod Deployment

Deploy directly to RunPod using the included configuration:

• View RUNPOD_DEPLOY.md for detailed instructions
• Includes automatic Redis setup and persistent storage

Documentation

• README.md - Setup and usage
• RUNPOD_DEPLOY.md - RunPod deployment guide
• .runpod/hub.json - Serverless configuration

---

Evolved from the Raspberry Pi agent project, now portable and cloud-ready