Model Providers

Model providers are the AI backends that power Ruri's conversations. You can connect to cloud providers like OpenAI, Anthropic, and Google Gemini, or run models locally on your own machine.

Provider Types

Ruri supports three provider types, covering almost all mainstream AI model services:

Type	Use Case	Common Services
OpenAI Compatible	The most universal type; most providers use this format	OpenAI, DeepSeek, Ollama, Zhipu, etc.
Anthropic Compatible	Designed for Anthropic's Claude series	Anthropic, Claude proxy services
Gemini Compatible	Google Gemini series models	Gemini 2.5 Flash, Gemini 2.5 Pro

提示

Not sure which one to choose? Go with OpenAI Compatible! It's the most universal format and works with the vast majority of providers.

Managing Providers

Adding a Provider

1. Navigate to the Providers page in the sidebar
2. Click Add Provider
3. Select the provider type (OpenAI Compatible / Anthropic Compatible / Gemini Compatible)
4. Fill in the name, API URL, API Key, model name, and whether it supports multimodal
5. Click Save

Fetching Available Models

Ruri can automatically fetch the list of available models from your provider's API:

1. Go to the Providers page
2. Click the Fetch Models button on a provider
3. Ruri queries the provider's API and returns all available model IDs
4. Pick the model you want from the list

This works with OpenAI-compatible and Anthropic-compatible APIs, making it easy to discover models without manually looking up IDs.

Activating Providers

Only one provider can be active at a time. In the provider list, click the Activate button on the provider you want to use.

提示

If you frequently switch between providers, check out Config Profiles — you can create profiles with different providers and switch between them instantly.

Editing and Removing

In the provider list, you can edit existing provider configurations or delete providers you no longer need at any time.

Popular Provider Setup

OpenAI (GPT-4o)

1. Add a provider and select OpenAI Compatible
2. Fill in the following:

Field	Value
Name	OpenAI
API URL	https://api.openai.com/v1
API Key	Your OpenAI API key (starts with sk-)
Model	gpt-4o
Supports Multimodal	true (GPT-4o supports image inputs)

3. Save and activate

信息

Get your API key from platform.openai.com. You'll need an OpenAI account with billing enabled.

Anthropic (Claude Sonnet 4)

1. Add a provider and select Anthropic Compatible
2. Fill in the following:

Field	Value
Name	Anthropic
API URL	https://api.anthropic.com
API Key	Your Anthropic API key
Model	claude-sonnet-4-20250514
Supports Multimodal	true (Claude supports image inputs)

3. Save and activate

信息

Get your API key from console.anthropic.com.

Google Gemini

1. Add a provider and select Gemini Compatible
2. Fill in the following:

Field	Value
Name	Gemini
API URL	https://generativelanguage.googleapis.com
API Key	Your Gemini API key
Model	gemini-2.5-flash
Supports Multimodal	true (Gemini natively supports image inputs)

3. Save and activate

信息

Get your API key from aistudio.google.com. Gemini offers a generous free tier — a great way to get started without paying!

Model Options

Popular Gemini models include:

• gemini-2.5-flash — Fast and efficient for most tasks
• gemini-2.5-pro — More capable for complex reasoning

DeepSeek

1. Add a provider and select OpenAI Compatible
2. Fill in the following:

Field	Value
Name	DeepSeek
API URL	https://api.deepseek.com/v1
API Key	Your DeepSeek API key
Model	deepseek-chat
Supports Multimodal	false

3. Save and activate

信息

DeepSeek is compatible with the OpenAI API format, so choose OpenAI Compatible. Get your API key from platform.deepseek.com.

Ollama (Local, Free)

1. First, install and start Ollama, then pull a model (e.g., ollama pull llama3)
2. Add a provider in Ruri and select OpenAI Compatible
3. Fill in the following:

Field	Value
Name	Ollama
API URL	http://localhost:11434/v1
API Key	Anything (e.g., ollama) — Ollama doesn't require a real key
Model	The model you pulled (e.g., llama3)
Supports Multimodal	false (unless using a vision model like llava)

4. Save and activate

提示

Ollama runs entirely locally — no payment required and no internet connection needed. If you're not ready to purchase API services yet, start with Ollama to experience all of Ruri's features!

Multimodal Support

Some model providers support multimodal inputs — the ability to include images in your chat messages. When enabled, you can send images to the AI and it can "see" and analyze them.

How It Works

• If your provider has Supports Multimodal set to true, the AI can receive images you paste or attach in the chat
• If a model receives an image but doesn't support it, Ruri automatically falls back by stripping the image and retrying the request — no errors, just graceful degradation
• Gemini and Claude models natively support multimodal, while many OpenAI-compatible models (GPT-4o, GPT-4V) also support it

Setting Multimodal Support

When adding or editing a provider, you can toggle the Supports Multimodal option:

• On — The provider will receive image content in requests
• Off — Images are stripped before sending to the provider (useful for local models that don't support vision)

警告

If you enable multimodal for a model that doesn't support it, you may see errors. Ruri handles this gracefully by retrying without images, but it's better to set this accurately from the start.

Tips

• 💡 Try Ollama first — Free local models, perfect for experimenting and testing, no API key needed
• 🆓 Try Gemini's free tier — Google Gemini offers a generous free tier, great for starting at zero cost
• 🔄 Set up multiple providers — Add multiple providers and switch between them anytime to handle different scenarios
• 📋 Config profile linking — Bind different providers in Config Profiles to switch your entire configuration with one click
• 🔗 Proxy services — If you use an API proxy (such as a relay service), simply change the API URL to your proxy endpoint
• ✅ Check URL paths — OpenAI-compatible URLs typically end with /v1, while Anthropic-compatible URLs do not
• 🔍 Fetch models — Use the "Fetch Models" button to automatically discover available models from your provider's API