novelai-openclaw-adaptor 0.1.1

NovelAI shim and image CLI for OpenClaw-friendly usage.

🎨 NovelAI OpenClaw Adaptor

A seamless shim to connect NovelAI's API to OpenClaw using the OpenAI format.

"Solving the core pain point of integrating NovelAI with OpenClaw by providing a simple, OpenAI-compatible local proxy."

Quick Start · How It Works · 简体中文 · 日本語

Why NovelAI OpenClaw Adaptor? 🌟

The core pain point for many users is that NovelAI's API cannot be directly integrated into OpenClaw, as OpenClaw expects the standard OpenAI API format.

To solve this, we created a shim (adaptor) that acts as a middleman. It translates OpenAI-formatted requests from OpenClaw into NovelAI-compatible requests, and returns the generated images seamlessly.

Additionally, it supports a simplified generation method: you can just pass your prompt directly to generate images without complex configurations!

Quick Start 🚀

📥 Install:

pip install novelai-openclaw-adaptor

⚙️ Initialize config:

novelai-config init

▶️ Start the text model shim:

novelai-shim

🖼️ Generate an image:

novelai-image --prompt "1girl, solo, masterpiece, best quality"

❓ Help:

novelai-config --help
novelai-shim --help
novelai-image --help

Connect to OpenClaw:

Tell OpenClaw directly:

Help me install novelai-openclaw-adaptor and add a new model provider "novelai": pip install novelai-openclaw-adaptor && novelai-config --help

Supported Models

Text models exposed by the shim:

• glm-4-6
• erato
• kayra
• clio
• krake
• euterpe
• sigurd
• genji
• snek

Image models supported by the image CLI:

• nai-diffusion-4-5-full
• nai-diffusion-4-5-curated
• nai-diffusion-4-full
• nai-diffusion-4-curated
• nai-diffusion-3
• nai-diffusion-3-furry

How It Works ✨

1. Local Proxy: The adaptor runs locally and provides an OpenAI-compatible endpoint (e.g., http://localhost:xxxx/v1).
2. Format Translation: When OpenClaw sends an OpenAI-style request, the adaptor translates it into NovelAI's specific parameters.
3. Simple Prompting: You can simply pass the prompt string to generate images.
4. Configuration: In OpenClaw, set the base_url to your local adaptor address, for example http://127.0.0.1:11434/v1, and configure the model name as required.

License 📄

This project is released under the MIT license.