Ryokan

AI Providers & Models

Learn which AI providers Ryokan supports, how to set up API keys, and how to use free, paid, or locally hosted AI models.

AI Providers & Models

Ryokan uses AI models to continue your stories. You decide which model you want to use, whether through a cloud provider or directly on your own device.

Supported AI Providers

Ryokan supports several cloud providers:

  • OpenRouter – Access to a wide range of models from different developers.
  • NanoGPT – Various AI models through a single API.
  • OpenAI – Direct access to OpenAI models.
  • Anthropic – Claude models.
  • Google Gemini – Google's Gemini models.
  • xAI / Grok – AI models from xAI.

You can also connect your own OpenAI-compatible API endpoints.

Setting Up an API Key

To use cloud models, you'll usually need an API key. You can get one from your chosen provider's website.

In Ryokan, go to Settings → API Connection, select your provider, and enter your key using Add API Key.

Ryokan will then automatically load the available models. Select one to use it for your stories.

When you first launch Ryokan, the setup wizard will guide you through these steps.

Important: Keep your API key private. Never share it with others.

Models and Pricing

Ryokan itself is free. Whether using an AI model costs money depends on the provider and model you choose.

Some providers offer free models, while others charge based on token usage. Tokens are small units of text that make up your inputs and the AI's generated responses.

Ryokan's model picker includes several useful features:

  • Search: Quickly find specific models.
  • Free: Filter available free models.
  • Favorites: Save models you use frequently.
  • Recent: Access recently used models.
  • Pricing: Compare displayed model prices when pricing information is available.

Free models may have usage limits. Prices and availability are determined by the provider and may change over time.

Local AI Models

You can also use Ryokan without a cloud provider. Supported options include:

  • LM Studio
  • Ollama
  • llama.cpp
  • KoboldCPP

To get started, launch a local model server and load your preferred model. Then select the corresponding provider in Ryokan.

Common server addresses are already configured. You can also adjust the API endpoint if needed.

Local models usually don't require an API key and don't incur usage-based cloud API fees. However, your device needs sufficient computing power to run them.

Using Multiple Connections

Ryokan lets you create multiple connection profiles and switch between them.

For example, you can use different cloud providers or switch to your local model at any time without having to configure everything again.

Try different models to discover which writing style works best for your stories.

Tip: All you need for your first story is a connected provider and a selected model. You can explore advanced settings later.