Corelia — Enterprise AI Orchestrator

🌐 GitHub Repo

Overview

Corelia is an enterprise backend orchestrator for AI agents (Finance, HR, IT, Security), each wrapping a local transformer model (Ollama/llama.cpp). Agents collaborate via a message bus, store context in a vector DB, and execute configurable workflows. Includes secure authentication, agent APIs, workflow runner, enterprise integrations, and usage examples.

📚 Documentation

🗺️ Architecture Diagram

flowchart TD
	subgraph API
		A1[FastAPI Backend]
		A2[JWT Auth & RBAC]
	end
	subgraph Agents
		B1[FinanceAgent]
		B2[HRAgent]
		B3[ITAgent]
		B4[SecurityAgent]
	end
	subgraph Integrations
		C1[Slack]
		C2[Jira]
		C3[GitHub]
		C4[Confluence]
	end
	subgraph Memory
		D1[Qdrant/Weaviate]
		D2[In-Memory Cache]
	end
	subgraph Infra
		E1[Redis Pub/Sub]
		E2[Postgres/SQLite]
		E3[Ollama Server]
	end
	A1 --> B1
	A1 --> B2
	A1 --> B3
	A1 --> B4
	B1 -- Ollama --> E3
	B2 -- Ollama --> E3
	B3 -- Ollama --> E3
	B4 -- Ollama --> E3
	B1 -- Memory --> D1
	B2 -- Memory --> D1
	B3 -- Memory --> D1
	B4 -- Memory --> D1
	B1 -- Cache --> D2
	B2 -- Cache --> D2
	B3 -- Cache --> D2
	B4 -- Cache --> D2
	B1 -- Pub/Sub --> E1
	B2 -- Pub/Sub --> E1
	B3 -- Pub/Sub --> E1
	B4 -- Pub/Sub --> E1
	B1 -- Tools --> C1
	B1 -- Tools --> C2
	B1 -- Tools --> C3
	B1 -- Tools --> C4
	B2 -- Tools --> C1
	B2 -- Tools --> C2
	B2 -- Tools --> C3
	B2 -- Tools --> C4
	B3 -- Tools --> C1
	B3 -- Tools --> C2
	B3 -- Tools --> C3
	B3 -- Tools --> C4
	B4 -- Tools --> C1
	B4 -- Tools --> C2
	B4 -- Tools --> C3
	B4 -- Tools --> C4
	A1 --> B1
	A1 --> B2
	A1 --> B3
	A1 --> B4
	A1 -- Logs --> E2

Loading

• Agent Communication: Redis pub/sub
• Memory: Qdrant/Weaviate + in-memory cache
• Workflows: YAML/JSON runner
• Integrations: Slack, Jira, GitHub, Confluence
• Persistence: PostgreSQL/SQLite logs

---

🚀 Quickstart

1. Clone & Install

git clone https://github.com/UnexpectedNull/Corelia.git
cd Corelia
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

2. Configure

Edit .env for Ollama, DB, Redis, Qdrant, and OpenLlori keys:

OLLAMA_SERVER_URL=http://localhost:11434
API_KEYS=yourkey1,yourkey2
PORT=8000
RATE_LIMIT=60

3. Start Services

docker-compose up -d

4. Run API

uvicorn main:app --host 0.0.0.0 --port 8000 --reload

5. Try Examples

cd examples
python query_agent.py
python run_workflow.py

---

⚡ OpenLlori Quickstart

Install dependencies:

pip install -r requirements.txt

Configure .env:

OLLAMA_SERVER_URL=http://localhost:11434
API_KEYS=yourkey1,yourkey2
PORT=8000
RATE_LIMIT=60

Start the server:

uvicorn main:app --host 0.0.0.0 --port 8000 --reload

Call the API (Python example):

import openai
openai.api_key = "yourkey1"
openai.api_base = "http://localhost:8000/v1"
response = openai.ChatCompletion.create(
	model="llama3",
	messages=[{"role": "user", "content": "Hello!"}]
)
print(response)

---

🗂️ Project Structure

main.py                # FastAPI entrypoint, routes
auth.py                # JWT + role-based access
agents/                # Agent logic (base + specializations)
workflows/             # Workflow runner + YAMLs
integrations/          # Slack, Jira, GitHub, Confluence
memory/                # Vector DB + cache
bus.py                 # Redis pub/sub
config.py              # Env loader
db.py                  # Logging (Postgres/SQLite)
requirements.txt       # Backend deps
docker-compose.yml     # Redis, Postgres, Qdrant
examples/              # Usage scripts

---

📝 Notes

• Agents use real Ollama and API integrations. Configure .env for all service tokens and endpoints.
• For production, use PostgreSQL and secure secrets.
• Extend workflows and integrations as needed.