Skip to content

README.md

Latest commit

 

History

History
202 lines (136 loc) · 5.54 KB

README.md

File metadata and controls

202 lines (136 loc) · 5.54 KB

🌶️ Spicychat API (Unofficial)

Python Version Async Status

A robust, asynchronous, and fully-typed Python wrapper for the spicychat.ai API. Designed for developers who want to build bots, automate interactions, or integrate Spicychat characters into their own applications with zero friction.

✨ Key Features

  • 🧠 Auto-Persistent Memory: Chat sessions automatically detect existing conversations and load history. No more managing IDs manually.
  • ⚡ Dual-Layer Architecture:
    • High-Level: ChatSession for an easy, stateful, "human-like" coding experience.
    • Low-Level: SpicyClient for raw access to endpoints, search, and bulk operations.
  • ⏪ Smart Context Control: Built-in undo(), edit_last_user_message(), and regenerate() that keep local state and server state perfectly synced.
  • 🎭 Persona Management: Create, switch, and delete user personas programmatically.
  • 🎬 Director Mode: Full support for Director Mode commands and custom generation settings (Temperature, Top-K, etc.).
  • 🚀 Premium Model Support: Native support for DeepSeek R1, TheSpice, and other premium models.

📦 Installation

You can install the library directly from GitHub using pip. No manual cloning required.

pip install git+https://github.com/SSL-ACTX/spicychat-api.git

Requires git to be installed on your system

🚀 Quick Start

This example logs you in, finds a character, and starts a persistent chat session. If you close the script and run it again, it will remember where you left off.

import asyncio
from spicy import SpicyClient, ChatModel

async def main():
    client = SpicyClient()

    # 1. Login (Interactive OTP input on first run, auto-saved afterwards)
    await client.login("your_email@example.com")

    # 2. Start Chat (Auto-searches & Auto-resumes)
    session = await client.start_chat("Asuka Langley")

    # 3. Configure Model (Optional)
    session.model = ChatModel.DEEPSEEK_R1 

    # 4. Chat!
    print(f"Chatting with: {session.character.name}")
    response = await session.send("Hello there!")
    print(f"Response: {response.content}")

    await client.close()

if __name__ == "__main__":
    asyncio.run(main())

🏗️ Architecture

The library is built on a two-tier abstraction model to balance ease of use with raw power.

graph TD
    User([User Script])
    
    subgraph "High Level (Easy Mode)"
        Session[ChatSession]
        History[History Manager]
        Settings[GenerationSettings]
    end
    
    subgraph "Low Level (Power User)"
        Client[SpicyClient]
        Auth[AuthManager]
        HTTP[HttpManager]
    end
    
    API[Spicychat API]
    
    User -->|Start Chat| Session
    Session -->|Sync| History
    Session -->|Config| Settings
    Session -->|Wraps| Client
    Client -->|Auth| Auth
    Client -->|Requests| HTTP
    HTTP -->|HTTPS| API
Loading

🛠️ Advanced Usage

1. Controlling Conversation Flow (Undo & Edit)

Mistakes happen. The library handles the complex logic of deleting/editing message pairs (User + Bot) to keep the conversation clean.

# Undo the last interaction (Deletes BOTH the Bot's reply AND your last message)
await session.undo()

# Edit your last message (e.g., fixing a typo)
await session.edit_last_user_message("I actually meant to say this...")

# Edit the Bot's last message (e.g., fixing formatting)
await session.edit_last_bot_message("*She smiles warmly* Of course!")

# Regenerate the last response with different settings
await session.regenerate(temperature=1.2)

2. Generation Settings

You can control the creativity of the AI globally or per-message.

from spicy import GenerationSettings

# Global Session Settings
session.settings = GenerationSettings(
    temperature=0.7,
    max_new_tokens=200,
    top_p=0.9
)

# Per-Message Override (High creativity for just this prompt)
await session.send(
    "Tell me a crazy sci-fi story", 
    temperature=1.4, 
    max_new_tokens=400
)

3. Persona Management

Switch between different "masks" instantly.

# List all personas
personas = await client.get_personas()

# Switch active persona for this chat
await session.switch_persona("Iris Seravelle")

# Create a new persona programmatically
await client.create_persona(
    name="Dark Alter Ego",
    description="A villainous version of myself.",
    avatar_path="./avatars/evil_version.png"
)

4. Low-Level API Access

Need raw data? Bypass the ChatSession and use SpicyClient directly.

# Search for characters
results = await client.search("Nier Automata")
for hit in results.hits:
    print(hit.document.name, hit.document.id)

# Fetch raw conversation objects
convs = await client.get_conversations(limit=10)

# Rate a bot
await client.rate_bot("2B", action=RatingAction.LIKE)

⚙️ Configuration

  • Tokens: Authentication tokens are stored securely in ~/.config/spicychat-api/tokens.json.
  • Logging: By default, noisy logs (httpx) are suppressed. To enable debug logging:
import logging
logging.getLogger("spicy").setLevel(logging.DEBUG)

⚖️ Disclaimer

This library is an unofficial wrapper and is not affiliated with, endorsed by, or connected to Spicychat.ai. Use responsibly and in accordance with the platform's Terms of Service.