Remote Models
Remote models allow you to connect any OpenAI-compatible inference API to Menatic AI, with API keys stored securely server-side.
Adding a Remote Model
Section titled “Adding a Remote Model”- Navigate to AI Models
- Click Add Model
- Select Remote
- Fill in the form:
| Field | Description |
|---|---|
| Name | Display name (e.g., GPT-4o, Claude 3.5) |
| API Endpoint | Base URL of the OpenAI-compatible API |
| API Key | Your API key (stored encrypted in the database) |
| Model ID | Model identifier sent in API requests (e.g., gpt-4o) |
- Click Save
Provider Examples
Section titled “Provider Examples”OpenAI
Section titled “OpenAI”Endpoint: https://api.openai.com/v1Model ID: gpt-4oAzure OpenAI
Section titled “Azure OpenAI”Endpoint: https://<resource>.openai.azure.com/openai/deployments/<deployment>Model ID: gpt-4oAnthropic (via proxy)
Section titled “Anthropic (via proxy)”Anthropic’s native API is not OpenAI-compatible, but you can use a proxy like LiteLLM to adapt it.
Self-Hosted (vLLM, llama.cpp)
Section titled “Self-Hosted (vLLM, llama.cpp)”Any server that implements the OpenAI /v1/chat/completions endpoint works as a remote model.
Using Remote Models
Section titled “Using Remote Models”Remote models appear in the sidebar under the Remote section alongside local Ollama models. They can be selected in the Admin Chat assistant and any other interface that supports model selection.
Security
Section titled “Security”API keys are stored encrypted in the platform database. They are never exposed in the frontend or logs. Requests to remote providers are made server-side by the backend.