Model Service Configuration
BLACK HOLE ships with connection templates for 60+ providers (model service vendors), covering most mainstream models from around the world as well as local inference frameworks. Use the guides below for provider-specific configuration, or the quick reference to find other providers.
Find a Provider
Open the quick reference to find a provider-specific guide.
Understand Connection Settings
Learn about API keys, API hosts, and advanced configuration.
Provider Types
BLACK HOLE groups providers by protocol into the following types, which behave slightly differently:
| Type | Compatible protocol | Typical examples |
|---|---|---|
| OpenAI-compatible | /v1/chat/completions | OpenAI, DeepSeek, SiliconFlow, OpenRouter, most third-party gateways |
| Anthropic-compatible | /v1/messages | Anthropic, BLACK HOLE IN, some gateways. BLACK HOLE Agent requires this type |
| Gemini | Google AI Studio / Vertex | Google Gemini, Vertex AI |
| Bedrock | AWS Bedrock SDK | AWS Bedrock |
| Azure OpenAI | Azure OpenAI Service | Azure OpenAI |
| Local inference | Local HTTP service | Ollama, LM Studio, GPUStack, OpenVINO Model Server |
| Special gateways | Vendor-specific protocols | NewAPI, OneAPI, AiHubMix, DMXAPI, etc. |
General Steps to Add a Provider
Open Settings
Open Settings → Model Provider
Select a Provider
Find the provider in the built-in list and click it to open its details page
Enter Connection Details
Enter the API Key (required) and change the API Host if needed (it defaults to the provider's official address)
Add Models
Click Sync models and add the chat, embedding or vision models you use
Check the Connection (Optional)
(Optional) Click Model Check to verify the connection with any chat model
Provider Guides
General / Gateways
- BLACK HOLE (Free)
- BLACK HOLE IN — dual endpoints (OpenAI + Anthropic), recommended for BLACK HOLE Agent
International Providers
- OpenAI
- Anthropic (Claude)
- Azure OpenAI
- Grok
- Groq
- OpenRouter
- Google Gemini
- Mistral
- Perplexity
- GitHub Copilot
- MiniMax M Plan
Chinese Providers
Local Inference
Custom Providers
- Custom Provider — any OpenAI / Anthropic / Gemini-compatible endpoint
Can't find the provider you use?
Look in the All Providers Quick Reference first. Open the dedicated guide where one is available; otherwise use the general steps on this page and Model Service Settings. For a compatible endpoint, see Custom Provider.
API Key and API Host
See Model Service Settings for details, including advanced usage such as rotating multiple keys and ending the host with # to use a fixed path.
Get Help and Submit Feedback
If you have any questions, bugs, or feature suggestions during configuration or use, please use the official channels listed in Feedback and Suggestions.