3.3 Supported AI Service Providers
MiniTavern supports multiple mainstream AI service providers. You can choose according to your needs.
On the Configure LLM page, available providers include: Nvidia, OpenAI, Anthropic, Google, Deepseek, Tencent, OpenRouter, Other LLM.
Nvidia
Representative Models: Llama, Mistral and other open-source models
Features:
- High free quota
- Supports multiple open-source models
How to Get:
- Official Website: https://build.nvidia.com/
- Register to get free quota: Nvidia Free API Setup
OpenAI
Representative Models: GPT-5, GPT-4 Turbo
Features:
- Industry-leading conversation capabilities
- Supports ultra-long context (128K tokens)
Use Cases:
- Need strongest conversation capabilities
- Complex role-playing
- Long conversations and story creation
- Official NSFW content support
How to Get:
- Official Website: https://platform.openai.com
- Pay-as-you-go
Anthropic (Claude)
Representative Models: Claude 4.5 Sonnet
Features:
- Ultra-long context (200K tokens)
- Powerful long-text understanding capabilities
Use Cases:
- Need to process long texts
- Value safety and ethics
- Complex character settings
How to Get:
- Official Website: https://console.anthropic.com
- Pay-as-you-go
Google (Gemini)
Representative Models: Gemini 3 Pro
Features:
- Ultra-large context window (1M tokens)
- Powerful multimodal capabilities
- High free quota
- Fast response speed
Use Cases:
- Budget-conscious users
- Need ultra-long context
- Multilingual conversations
How to Get:
- Official Website: https://aistudio.google.com/app/api-keys
- Has free quota
- Pay-as-you-go
DeepSeek
Representative Models: DeepSeek-Chat, DeepSeek-Reasoner
Features:
- Open-source model
- Extremely high cost-effectiveness
- Powerful coding capabilities
Use Cases:
- Budget-sensitive users
- Daily conversations
How to Get:
- Official Website: https://platform.deepseek.com
Tencent (Hunyuan)
Representative Models: Hunyuan-Large, Hunyuan-Standard
Features:
- Tencent Cloud product
- Excellent Chinese language capabilities
How to Get:
- Official Website: https://cloud.tencent.com/product/hunyuan
OpenRouter
Representative Models: Aggregates multiple models (GPT-5, Claude, Llama, etc.)
Features:
- One API to access multiple models
- Unified billing, easy to manage
- Automatically selects optimal model
- Supports open-source models
Use Cases:
- Want to try various models
- Don't want to manage multiple accounts
- Need flexible switching
How to Get:
- Official Website: https://openrouter.ai
Other LLM (Custom/Ollama)
Supports Ollama and other OpenAI-compatible interfaces.
Ollama Local Deployment:
- Completely free
- Data completely localized
- No network connection required
- Official Website: https://ollama.ai
Other Custom Interfaces:
- Must be compatible with OpenAI API format
- Need to provide complete API URL
- Need valid API Key