KKR Mind — Unified AI Interface

One API. One private AI engine. 27 agents. Memory. Streaming. No model names.

KKR Mind routes every request through a single private AI engine. No vendor names. No model IDs. Just ask, and get an answer.

✨ Features

• One private engine — a single unified AI behind one simple API (no model names are ever exposed, by design)
• 27 Built-in Agents — fitness coach, coding mentor, business consultant, travel planner, and more
• Custom Agents — create your own AI personas, stored locally
• Memory — multi-turn conversations with context
• Streaming UX — word-by-word responses for a live feel
• Any language — Bengali, Hindi, English, and more
• Health Checks — ping() diagnostics with latency
• Privacy-First — local storage only, no cloud, no tracking

📦 Installation

pip install kkrmind

From source:

git clone <your-repo-url> kkr-mind
cd kkr-mind
pip install -e .

Requirements: Python 3.7+, requests>=2.28.0 (auto-installed).

🚀 Quick Start

from kkrmind import KKRMind

ai = KKRMind()

# Chat
response = ai.chat("Hello!")
print(response)

# Any language
print(ai.chat("তুমি কে?"))

# Streaming (word-by-word)
for chunk in ai.chat_stream("Tell me a short story"):
    print(chunk, end="", flush=True)

🧠 Memory

ai = KKRMind(memory=True, max_history=50)

ai.chat("My name is Rahul")
print(ai.chat("What is my name?"))   # -> "Your name is Rahul."

ai.clear_memory()

🤖 Built-in Agents (27)

print(ai.chat_with_agent("coding-mentor", "Explain decorators"))
print(ai.chat_with_agent("travel-planner", "Plan a 3-day trip to Kolkata"))

# Discover agents
for agent_id, info in ai.list_agents().items():
    print(agent_id, "-", info["name"])

print(ai.search_agents("finance"))

👤 Custom Agents

Create AI agents with custom personas. Stored locally (~/.kkrmind/custom_agents.json) for privacy.

agent_id = ai.create_agent(
    name="Python Expert",
    description="Senior Python developer",
    system_prompt="You are a senior Python developer with 20 years of experience...",
)

print(ai.chat_with_agent(agent_id, "Explain decorators"))

for a in ai.list_custom_agents():
    print(a["id"], a["name"])

ai.update_agent(agent_id, description="Updated!")
ai.delete_agent(agent_id)

🩺 Health Checks

report = ai.ping()
print(f"Status: {report['status']}, Latency: {report['latency_ms']}ms")

🔌 Backend

KKR Mind is built on top of a single private Cloudflare Worker:

GET https://img-thumb-cache.iamkkronly58.workers.dev/api
    ?chat=<message>
    &system=<optional persona>

{
  "status": "ok",
  "query": "hello",
  "response": "Hello! How can I help you today?",
  "latency_ms": 1523,
  "developer": "Kaustav Kanti Ray @iamkkronly"
}

The backend also reports a model field — KKR Mind discards it on purpose, so users never see model names. To point the library at a different backend, edit Config.API_BASE in kkrmind/config.py.

Image generation (coming soon)

The current backend is chat-first. The full image API (generate_image, AspectRatio, ImageStyle, ImageQuality) is already wired in and will activate as soon as an image endpoint is added to the backend Worker.

🔐 Privacy

• ✅ Local storage only — custom agents in ~/.kkrmind/
• ✅ No cloud sync — data never leaves your machine
• ✅ No telemetry — zero usage tracking
• ✅ No accounts — no registration required
• ✅ No model names — the engine stays private
• ✅ Full control — you own all your data

📁 Project Structure

kkr-mind/
├── LICENSE                  # MIT (with upstream credit)
├── MANIFEST.in
├── README.md
├── requirements.txt
├── setup.py                 # PyPI packaging
├── setup.cfg
├── BUILD_GUIDE.md           # How to build & publish this project
├── kkrmind/
│   ├── __init__.py          # Public API + help()
│   ├── client.py            # KKRMind client (chat, agents, ping, ...)
│   ├── config.py            # Backend endpoint + settings
│   ├── exceptions.py        # KKRMindError hierarchy
│   ├── models.py            # Single unified engine (no names)
│   ├── agents.py            # 27 built-in agents
│   ├── memory.py            # Conversation memory
│   ├── image.py             # GeneratedImage (for the future image API)
│   ├── utils.py             # Helpers
│   └── docs.md              # Full in-package documentation
└── examples/
    ├── demo_chat.py
    ├── demo_stream.py
    ├── demo_memory.py
    ├── demo_agents.py
    ├── demo_custom_agent.py
    └── demo_ping.py

🧩 Development

python -m venv .venv
source .venv/bin/activate        # Windows: .venv\Scripts\activate
pip install -e .
python examples/demo_chat.py

Full step-by-step (packaging, building the wheel, publishing to PyPI): see BUILD_GUIDE.md.

📚 In-package docs

import kkrmind
kkrmind.help()            # full docs
kkrmind.help("agents")    # one section

📄 License

MIT License — see LICENSE.

KKR Mind is an independent project by Kaustav Kanti Ray (@iamkkronly). Its architecture is inspired by / references AIMind by GlitchyAPI (pypi.org/project/aimind), also MIT licensed — thanks!

👤 Developer

কৌস্তভ কান্তি রায় (Kaustav Kanti Ray) — @iamkkronly

Built with ❤️ by Kaustav Kanti Ray