🚀 What I Built
I built Hermes Agent Assistant, a lightweight agentic AI system designed to demonstrate how modern AI agents can be structured using a modular architecture instead of a simple, single-prompt response model.
The system takes an abstract user task, breaks it down into structured steps using a dedicated planner, executes those steps sequentially via an execution engine, utilizes targeted tools, and stores the interaction context in a persistent memory system.
⚙️ Why I Built This
Most AI applications today are simple wrappers around LLMs that rely on a single input-output loop. I wanted to understand and demonstrate how production-grade, autonomous agent systems operate internally. Specifically, I wanted to explore how:
Planning can be decoupled from execution to allow for complex error handling and multi-step reasoning.
Tools can be dynamically integrated into an agent's reasoning loop.
State and memory can persist across tasks to enable true contextual continuity.
Hermes Agent is my architecture simulation built to solve this problem in a highly accessible, lightweight, and scalable format.
🧠 System Architecture & Workflow
The codebase is split cleanly into four autonomous components that mirror real-world AI agent meshes:
User Request (e.g., /run?task=...)
│
▼
┌───────────────────────────┐
│ PLANNER │ ➔ Slices abstract goals into
└─────────────┬─────────────┘ structured, sequential steps.
│
▼
┌───────────────────────────┐
│ EXECUTOR │ ➔ Orchestrates task completion
└─────────────┬─────────────┘ by processing each step.
│
▼
┌───────────────────────────┐
│ TOOLS LAYER │ ➔ Provides functional utilities
└─────────────┬─────────────┘ (simulated web search, logic, maths).
│
▼
┌───────────────────────────┐
│ MEMORY SYSTEM │ ➔ Persists execution logs statefully
└───────────────────────────┘ into local JSON storage.
📡 Production Showcases & Links
🌐 Live Production Demo:
🆔 Cloud Deployment Service ID:srv-d88revegvqtc73bdj380(Render Infrastructure Node)
💡 What Makes It Different
Unlike traditional, rigid APIs or simple conversational chatbots, Hermes Agent:
Thinks in Workflows: It establishes an internal chain-of-thought lifecycle before executing anything.
Separates Reasoning from Action: Slicing the Planner from the Executor prevents cascading generation failures.
Is Highly Extensible: New tools and custom utility logic can be dropped into the system without breaking core routing.
Maintains Context Persistence: The custom memory module ensures state history is preserved between network calls.
🎛️ API Interaction Example
Request
POST /run?task=search AI agents HTTP/1.1
Host: hermes-agent-tanush.onrender.com
Response
{
"task": "search AI agents",
"plan": [
"analyze request parameters",
"query tool registry for search utilities",
"summarize agent data structural output"
],
"result": "final structured output successfully generated and written to persistent storage."
}
🧰 Tech Stack
Core Language: Python 3.10+
Web Framework: FastAPI (Asynchronous Server Gateway Interface)
Production Server: Uvicorn
Memory Layer: Volatile-to-Persistent JSON state manager
Architecture Pattern: Modular Agentic Workflow Design
🔮 Future Improvements & Roadmap
- 🤖 Real Foundation LLM Integration: Swapping out simulated logic for live OpenAI, Anthropic, or local open-source Ollama completion hooks.
- 🗄️ Vector Database Memory Upgrade: Transitioning flatfile storage over to a proper semantic vector indexing framework (FAISS / ChromaDB) for semantic chunk lookups.
- 🤝 Multi-Agent Orchestration: Upgrading the workflow to host distinct
Planner,Executor, andCriticagents working collaboratively with separate system prompts. - ⚡ Live Server-Sent Events (SSE): Integrating real-time execution streaming so client frontends can observe the agent's thought process step-by-step.
SOCIAL SHARE CARD GENERATOR