chalks 0.1.0

A minimalist real-time messaging platform

A minimalist real-time messaging service.

Chalk is a lightweight messaging service built on WebSocket. It provides group chats, direct messages, and real-time event delivery - nothing more, nothing less.

---

What is Chalk?

Chalk is a messaging service, Think of it as a simple, self-hosted alternative to Slack or Discord, designed for programmatic use.

What Chalk provides:

• Real-time messaging via WebSocket
• Group chats and direct messages
• Message history and persistence
• Event-driven message handling
• User authentication

What Chalk does NOT provide:

• AI/LLM capabilities
• Agent orchestration
• Task scheduling
• Business logic

Chalk is infrastructure. You build whatever you want on top of it.

---

Core Features

Minimalist API

Start a server in 2 lines, connect a client in 3:

# Server
from chalk.server import ChalkServer

server = ChalkServer(
    redis_url="redis://localhost:6379",  # Redis for pub/sub
    db_path="chalk.db",                   # SQLite for data persistence
    host="0.0.0.0",                       # Listen on all interfaces
    port=8000                              # Server port
)
server.run()

# Client
from chalk.client import Client

alice = Client(
    name="alice",
    password="password123",
    server="localhost:8000"  # Server address
)
chat = await alice.create_group_chat("My Chat")
await chat.send("Hello!")

Real-Time Messaging

• WebSocket-based: Instant bidirectional communication
• Event-driven: React to messages with decorators
• Persistent: Full message history in SQLite
• Reliable: Auto-reconnect on network issues

Group & Direct Chats

• Group chats: Multi-party conversations, unlimited members
• Direct messages: 1-on-1 private chats
• Message history: Query past messages anytime
• Member management: Add/remove members dynamically

---

Quick Start

Installation

pip install chalks

1. Start the Server

from chalk.server import ChalkServer

server = ChalkServer(
    redis_url="redis://localhost:6379",  # Redis for pub/sub
    db_path="chalk.db"                    # SQLite for data persistence
)
server.run()

Prerequisites:

• Redis must be running: redis-server

2. Send Your First Message

import asyncio
from chalk.client import Client

async def main():
    # Create client
    alice = Client("alice", "password123")
    
    # Create a group chat
    chat = await alice.create_group_chat("My First Chat")
    await chat.send("Hello world!")
    
    print(f"Chat created: {chat.id}")

asyncio.run(main())

3. Receive Messages

# Bob joins the same chat
bob = Client("bob", "password456")

# Register message handler
@bob.on("message")
async def handle_message(msg):
    print(f"{msg.sender.name}: {msg.content}")

# Join Alice's chat
await bob.join_chat(chat.id)

# Bob receives all messages in real-time
await asyncio.Event().wait()

---

Core Concepts

Client

Represents a user or agent in the system:

alice = Client(
    name="alice",           # Username
    password="pass123",     # Password (auto-registers if new)
    bio="AI assistant",     # Optional bio
    server="localhost:8000"  # Server address
)

Chat

Two types of chats:

Group Chat - Multi-party conversations:

# Create group
chat = await alice.create_group_chat("Team Chat")

# Add members
await chat.add_member(bob_id)
await chat.add_member(charlie_id)

# Multiple members can join

Direct Chat - 1-on-1 conversations:

# Create direct chat with Bob
dm = await alice.create_direct_chat(bob_id)

Message

Messages with rich metadata:

@client.on("message")
async def handle(msg):
    print(f"From: {msg.sender.name}")
    print(f"Content: {msg.content}")
    print(f"Time: {msg.timestamp}")
    
    # Reply to message
    await msg.reply("Got it!")
    
    # Get chat context
    chat = await msg.get_chat()
    members = await chat.get_members()

---

Common Patterns

Chat Management

# List all chats
chats = await client.list_chats()

# Filter by type
group_chats = [c for c in chats if c.is_group()]
direct_chats = [c for c in chats if c.is_direct()]

# Get chat details
chat = await client.get_chat(chat_id)
members = await chat.get_members()
messages = await chat.get_messages(limit=50)

# Join/leave
await client.join_chat(chat_id)
await client.leave_chat(chat_id)

---

API Reference

Client Methods

Method	Description
create_group_chat(name, members=[])	Create a group chat
create_direct_chat(user_id)	Create/get 1-on-1 chat
list_chats()	List all chats
get_chat(chat_id)	Get chat details
join_chat(chat_id)	Join a chat
whois(username)	Find users by name
stop()	Disconnect client

Chat Methods

Method	Description
send(content)	Send a message
get_messages(limit=50)	Get message history
get_members()	Get chat members
is_group()	Check if group chat
is_direct()	Check if direct chat

Message Methods

Method	Description
reply(content)	Reply to this message
get_chat()	Get the chat this message belongs to
get_sender()	Get sender details

---

Examples

comming soon...

---

Architecture

┌─────────────────┐         ┌─────────────────┐
│  Agent/Human    │         │  Agent/Human    │
│  (Client SDK)   │         │  (Client SDK)   │
└────────┬────────┘         └────────┬────────┘
         │                           │
         │      WebSocket + HTTP     │
         │                           │
         └───────────┬───────────────┘
                     │
         ┌───────────▼────────────┐
         │   Chalk Server         │
         │  (FastAPI + WebSocket) │
         └───────────┬────────────┘
                     │
         ┌───────────┴───────────┐
         │                       │
    ┌────▼─────┐          ┌─────▼────┐
    │  Redis   │          │ SQLite   │
    │ (Pub/Sub)│          │  (Data)  │
    └──────────┘          └──────────┘

Tech Stack:

• Server: FastAPI + Uvicorn
• Real-time: WebSocket + Redis Pub/Sub
• Storage: SQLite (Peewee ORM)
• Client: httpx + websockets
• Tasks: Huey (async task queue)

---

Is Chalk for You?

Use Chalk if you need:

• Real-time messaging infrastructure without the bloat
• WebSocket-based communication that's easy to integrate
• Self-hosted messaging (no third-party dependencies)
• Simple API that gets out of your way
• Group chats and DMs with message persistence

Chalk does one thing: messaging. It does it well.

---

License

MIT License - see LICENSE file for details.

---

Contributing

Contributions welcome! Please open an issue or PR.

---

Built with ❤️ for the AI agent community