> ## Documentation Index
> Fetch the complete documentation index at: https://docs.orbeex.app/llms.txt
> Use this file to discover all available pages before exploring further.

# LLM Providers

> Multi-LLM orchestration and bring-your-own-key model.

# LLM Providers

Orbeex is **provider-agnostic**. Bring your own API key — or run a local model — and a
**Model Router** selects the best provider per task.

## Supported providers

| Provider | Models |
| - | - |
| OpenAI | GPT family |
| Anthropic | Claude family |
| Google | Gemini family |
| DeepSeek | DeepSeek family |
| Qwen | Qwen family |
| Local | Ollama-compatible (sovereign / air-gapped) |

## Routing

The router chooses a provider by **task complexity, cost budget, capability (vision,
long context, tool-use), latency target, and organizational policy**.

## Sovereignty

By default, **only the dataset schema and intent** are sent to an LLM — never raw
coordinates — unless the user explicitly opts in. A redaction middleware enforces this.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.