MyOwnChat Python SDK & AI Agent Messaging Gateway

Zero-Configuration Messaging Bridge for AI Agents, Hermes Agent, and Chat Bots

Connect any AI Agent to MyOwnChat using the same 1-key simplicity as Telegram (@BotFather) and Slack (Socket Mode).

---

Key Features

• ⚡ 1-Key Setup: No channel_id, user_id, or database setup needed. LittleBackend Gateway Auto-Scoping automatically binds all reads and writes to the App Channel.
• 📡 Real-Time SSE Stream: Instant message delivery via Server-Sent Events with automatic cursor tracking and reconnects.
• 🧠 Universal OpenAI-Compatible Bridge: Connect to any OpenAI-compatible LLM endpoint — including local LLMs (LM Studio, Ollama, llama.cpp, vLLM) and cloud providers (OpenAI, OpenRouter, Groq, Together AI, DeepSeek).
• 💬 Live Stream Interaction: Instant thinking indicators and periodic 5-second line progress snippets while generating.
• 🔄 Channel Context Memory: Reconstructs multi-turn history from channel messages upon boot, preserving context across gateway restarts.
• 🛡️ Loop Prevention: Built-in filtering to ignore bot echoes, webhooks, and system notices.
• 🧵 Threading & Formatting: Full support for Markdown, emojis, and reply_to_id message threading.
• 🏛️ Hermes Agent Ready: Works out of the box with Nous Research Hermes Agent.

---

Installation

pip install myownchat
# or for local development:
pip install -e .

---

Quickstart (3 Lines of Code)

from myownchat import MyOwnChatGateway

gateway = MyOwnChatGateway(api_key="lb_live_...")


@gateway.on_message
def handle_message(msg):
    return f"Hello! You said: {msg.text}"


gateway.start()

---

Universal OpenAI-Compatible Bridge

Run the gateway connected directly to your local or cloud LLM:

1. LM Studio (Local)

export MYOWNCHAT_API_KEY="lb_live_..."
export OPENAI_BASE_URL="http://localhost:1234/v1"
export OPENAI_MODEL="local-model"

python -m myownchat --mode openai

2. Ollama (Local)

export MYOWNCHAT_API_KEY="lb_live_..."
export OPENAI_BASE_URL="http://localhost:11434/v1"
export OPENAI_MODEL="llama3.2"

python -m myownchat --mode openai

3. llama.cpp Server (Local)

export MYOWNCHAT_API_KEY="lb_live_..."
export OPENAI_BASE_URL="http://localhost:8080/v1"

python -m myownchat --mode openai

4. OpenAI (Cloud)

export MYOWNCHAT_API_KEY="lb_live_..."
export OPENAI_API_KEY="sk-..."
export OPENAI_MODEL="gpt-4o-mini"

python -m myownchat --mode openai

---

Hermes Agent Integration

To connect Nous Research Hermes Agent to MyOwnChat:

export MYOWNCHAT_API_KEY="lb_live_..."
export HERMES_API_KEY="your_api_key"
python examples/hermes_agent_bridge.py

---

Direct Message Posting via CLI

Send a one-off notification or message directly to your channel without running a persistent daemon:

export MYOWNCHAT_API_KEY="lb_live_..."
python -m myownchat --send "Deployment completed successfully! 🚀"

Or pass the key inline:

python -m myownchat -k "lb_live_..." -s "Hello from the CLI"

---

Code Examples

1. Async Message Handler

import asyncio
from myownchat import MyOwnChatGateway

gateway = MyOwnChatGateway(api_key="lb_live_...")


@gateway.on_message
async def handle_message(msg, gw):
    await asyncio.sleep(0.5)  # Perform async tasks or API calls
    return f"Processed asynchronously: {msg.text}"


gateway.start()

2. Low-Level REST Client

from myownchat import MyOwnChatClient

client = MyOwnChatClient(api_key="lb_live_...")

# Fetch channel metadata
channel = client.get_channel()
print(f"Connected to #{channel.name} (ID: {channel.id})")

# Post a message
msg = client.post_message("Hello from the REST client!")

# Fetch recent message history (auto-scoped to your channel)
history = client.get_messages(limit=10)
for m in history:
    print(f"[{m.id}] {m.text}")

3. Threaded Replies & Progress Editing

import time
from myownchat import MyOwnChatGateway

gateway = MyOwnChatGateway(api_key="lb_live_...")


@gateway.on_message
def handle_task(msg, gw):
    # 1. Send initial status replying to the user
    status = gw.send("⏳ *Starting analysis...*", reply_to=msg)

    # 2. Update progress
    time.sleep(2)
    try:
        gw.edit(status, "📊 *Processing results (50%)...*")
    except Exception:
        pass

    # 3. Final completion
    time.sleep(2)
    try:
        gw.edit(status, "✅ *Analysis complete!* All checks passed.")
    except Exception:
        gw.send("✅ *Analysis complete!*", reply_to=msg)


gateway.start()

4. Rich Embed Cards & Custom Metadata

from myownchat import MyOwnChatClient

client = MyOwnChatClient(api_key="lb_live_...")

client.post_message(
    text="Deployment Report 📋",
    metadata={
        "embed": {
            "title": "Production Deployment",
            "description": "Release **v2.4.0** deployed successfully.",
            "color": "#10B981",
            "fields": [
                {"name": "Environment", "value": "Production", "inline": True},
                {"name": "Duration", "value": "42s", "inline": True},
            ],
        }
    },
)

---

Environment Variables

Variable	Description	Default
MYOWNCHAT_API_KEY	Machine App API Key (lb_live_...)	Required
MYOWNCHAT_API_BASE	MyOwnChat API Base URL	https://api.marcvali.org/myownchat
MYOWNCHAT_BOT_NAME	Display name of the bot	AI Assistant
OPENAI_BASE_URL	OpenAI-compatible endpoint URL	https://api.openai.com/v1
OPENAI_API_KEY	API Key (or dummy for local LLMs)	not-needed
OPENAI_MODEL	Model identifier	gpt-4o-mini
OPENAI_SYSTEM_PROMPT	System prompt instruction	Friendly AI prompt
OPENAI_MAX_HISTORY	Max conversation turns retained	30
HERMES_API_KEY	Nous / OpenRouter API Key for Hermes	None

---

License

MIT License. Developed for the MyOwnChat ecosystem.