Orbicus LogoOrbicusBeta
Product

Model Provider Setup

Orbicus gives you complete freedom to connect to any cloud provider, custom API proxy, or self-hosted local model. You are never restricted to hardcoded model lists—simply enter your credentials and type or select any model ID.

1. Cloud Providers

Configure your API credentials in Settings across any Orbicus interface:

Anthropic

  • Setup: Provide your ANTHROPIC_API_KEY.
  • Supported Models: Use any Claude model ID (e.g. Claude Sonnet, Opus, or Haiku series with standard or extended thinking modes).

OpenAI

  • Setup: Provide your OPENAI_API_KEY.
  • Supported Models: Use any GPT flagship or reasoning model ID (e.g. o-series reasoning models, GPT-4o).

Google Gemini

  • Setup: Provide your GEMINI_API_KEY.
  • Supported Models: Use any Gemini Flash or Pro model ID.

2. Custom Compatible Endpoints

Orbicus supports custom API base URLs (baseUrl) across all major wire protocols. This enables you to connect directly to third-party model providers, inference aggregators, or internal corporate proxies.

Note: The endpoints and model IDs listed below are popular examples. You can connect any service or proxy that conforms to these wire protocols.

OpenAI-Compatible API Examples

Select the OpenAI provider type and specify your custom baseUrl:

  • DeepSeek: https://api.deepseek.com/v1 (Example models: deepseek-chat, deepseek-reasoner)
  • MiniMax: https://api.minimaxi.chat/v1 (Example models: MiniMax-Text-01)
  • OpenRouter: https://openrouter.ai/api/v1 (Example models: deepseek/deepseek-r1, anthropic/claude-3.7-sonnet)
  • Groq: https://api.groq.com/openai/v1 (Example models: llama-3.3-70b-versatile, deepseek-r1-distill-llama-70b)
  • Together AI: https://api.together.xyz/v1
  • Mistral AI: https://api.mistral.ai/v1 (Example models: mistral-large-latest, codestral-latest)
  • LiteLLM / Local AI Proxies: http://localhost:4000/v1

Anthropic-Compatible API Examples

Select the Anthropic provider type and supply your custom endpoint URL for services exposing the Anthropic Messages wire format:

  • OpenRouter (Anthropic Protocol): https://openrouter.ai/api/v1
  • MiniMax (Anthropic Protocol): https://api.minimaxi.chat/v1
  • Cloudflare AI Gateway: https://gateway.ai.cloudflare.com/v1/{account_id}/{gateway_id}/anthropic
  • AWS Bedrock / Enterprise Relays: https://your-enterprise-proxy.com/anthropic/v1
  • LiteLLM Anthropic Passthrough: http://localhost:4000/anthropic/v1

Google Gemini-Compatible API Examples

Select the Google provider type and enter your custom endpoint URL to route calls through custom Gemini proxies, regional gateways, or enterprise AI hubs:

  • Cloudflare AI Gateway for Gemini: https://gateway.ai.cloudflare.com/v1/{account_id}/{gateway_id}/google-ai-studio
  • Google Vertex AI Enterprise Proxies: https://your-custom-gateway.internal/gemini/v1beta
  • OpenRouter / Multi-Provider Hubs: https://openrouter.ai/api/v1
  • LiteLLM Google Passthrough: http://localhost:4000/gemini/v1

3. Local & Offline Models (Ollama, vLLM, LM Studio)

Run models completely offline with zero data leaving your machine:

Ollama Setup

  1. Install and start Ollama.
  2. Pull your desired model (example: deepseek-r1): 
    ollama run deepseek-r1
  3. In Orbicus Settings, select the OpenAI provider type and set:
    • Base URL: http://localhost:11434/v1
    • Model: deepseek-r1 (or your chosen model name)
    • API Key: ollama (placeholder)

vLLM / LM Studio

Point Orbicus to your local server endpoint (e.g. http://localhost:1234/v1).

Storage & Credential Security

All credentials and base URL endpoints are stored locally on your device in:

~/.orbicus/settings.json

API keys are never transmitted to Orbicus servers.