VectorFlow

AI-Powered Negotiation Intelligence System

---

---

Overview

VectorFlow is an AI-driven negotiation intelligence system that:

• Monitors incoming emails
• Analyzes negotiation context using an LLM-powered backend
• Generates structured negotiation insights
• Drafts high-quality strategic responses
• Automates workflows via n8n
• Runs fully containerized using Docker

---

Architecture Overview

flowchart LR
    A[Gmail Trigger] --> B[n8n Workflow]
    B --> C[FastAPI Backend]
    C --> D[Negotiation Agent]
    D --> E[Strategy Analysis + Draft Reply]
    E --> F[Gmail Send Node]
    F --> G[Mark as Read]

Loading

---

System Components

1. FastAPI Backend

• Handles negotiation analysis requests
• Structured response schema
• Modular agent architecture

2. Negotiation Agent

• Detects negotiation stage
• Assesses power balance
• Calculates win probability
• Estimates collapse risk
• Generates strategy options
• Produces draft reply

3. n8n Automation

• Gmail Trigger (Message Received)
• HTTP Request → Backend /analyze
• Send Response Email
• Mark Original Email as Read

4. Docker Infrastructure

• Fully containerized backend + automation
• Persistent n8n storage
• Auto-restart on system reboot

---

Request/Response Flow

sequenceDiagram
    participant Gmail
    participant n8n
    participant FastAPI
    participant Agent

    Gmail->>n8n: New Email
    n8n->>FastAPI: POST /analyze {thread}
    FastAPI->>Agent: Analyze Context
    Agent-->>FastAPI: Structured JSON
    FastAPI-->>n8n: Strategy + Draft Reply
    n8n->>Gmail: Send Analysis Email

Loading

---

Example Backend Response

{
  "recommended_strategy": "Value-based counter-anchor",
  "overall_win_probability": 0.72,
  "collapse_risk": 0.14,
  "draft_reply": "Thank you for the proposal..."
}

---

Project Structure

VectorFlow/
│
├── backend/
│   ├── app/
│   │   ├── agents/
│   │   │   └── negotiation_agent.py
│   │   ├── schemas/
│   │   │   └── negotiation.py
│   │   ├── main.py
│   │   └── __init__.py
│   ├── Dockerfile
│   ├── requirements.txt
│   └── .env
│
├── docker-compose.yml
└── start_vectorflow.bat

---

Local Deployment

Start System

docker compose up -d

Or simply run:

start_vectorflow.bat

Access:

• FastAPI Docs → http://localhost:8000/docs
• n8n UI → http://localhost:5678

---

Design Principles

• Modular agent-based architecture
• Clear separation between automation and intelligence layers
• Container-first deployment
• Resume-grade production structure
• Extensible to Slack, CRM, or enterprise pipelines

---

Future Improvements

• Human-in-the-loop approval node
• Vector database memory
• Conversation-level strategy tracking
• Cloud deployment (AWS/GCP/Azure)
• Multi-channel integration

---

License

MIT License

---

If this project helped you or inspired you, star this repo ⭐