Mixgarden Python SDK

pip install mixgarden

• Use get_models() to get a list of available models and providers.
• Use get_plugins() to get a list of available plugins.
• Use get_conversations() to get a list of available conversations.
• Use get_conversation(id) to get a specific conversation.
• Use chat(model, messages) when you’re building a conversational UI, or want the platform to maintain context for you.
• Use get_mg_completion(model, content) when you need a one-off, stateless generation enhanced by Mixgarden plugins.
• Use get_completion(model, messages) when you need a one-off, stateless generation directly from LLM (Raw LLM Output).

import os
from mixgarden import MixgardenSDK

sdk = MixgardenSDK(api_key=os.getenv("MIXGARDEN_API_KEY"))

# List available models
models = sdk.get_models()

# Stateful chat (conversation is stored server-side)
chat = sdk.chat(
    model="openai/gpt-4o-mini",
    # Optionally add conversationId to add to an existing conversation
    # conversationId: "your-convo-id",
    content="Hello mixgarden!",
    pluginId="tone-pro",
    pluginSettings={
        "emotion-type": "neutral",
        "emotion-intensity": 6,
        "personality-type": "friendly",
    },
)

# Other helper calls
plugins = sdk.get_plugins()
conversations = sdk.get_conversations()
conversation = sdk.get_conversation(conversations[0]["id"])

# One-shot completion (Mixgarden Enhanced)
completion = sdk.get_mg_completion(
    model="openai/gpt-4o-mini",
    content="hello mixgarden!",
    maxTokens=100,
    temperature=0.7,
    pluginId="tone-pro",
    pluginSettings={
        "emotion-type": "neutral",
        "emotion-intensity": 6,
        "personality-type": "friendly"
    }
)

# One-shot completion (Raw LLM Output)
completion = sdk.get_completion(
    model= "openai/gpt-4o-mini",
    messages=[{"role": "user", "content": "Hi there!"}],
    maxTokens=100,
    temperature=0.7,
)

print(chat)
print(completion)