Connections

Connect external model endpoints (OpenAI-compatible servers and Ollama instances) so their models can be discovered and used within your IntraLLM AI workspace.

Connections

The Connections page is the central hub for integrating external AI model endpoints into your IntraLLM AI workspace. Administrators can link endpoints (such as OpenAI-compatible servers or local Ollama instances) so their models can be accessed and used within the platform.

Where to find Connections

Open:

Admin Settings → Connections

Overview

The page is divided into two sections based on the type of API you are connecting:

  • OpenAI format API: Connect to any server that implements the OpenAI API standard.
  • Ollama API: Connect to Ollama instances for model discovery and chat requests.

Section 1: OpenAI format API

This section manages endpoints that are compatible with the OpenAI API structure.

Purpose

  • Manage OpenAI-compatible endpoints and their associated API keys for your workspace.

Configuration fields

  • Endpoint URL: The base URL of your OpenAI-compatible server (for example, http://192.168.18.49:8000/v1).
  • API Key: A secure field where the authentication key is stored. Keys are masked in the UI for privacy and safety.

Controls

  • Toggle switch: Enables or disables the specific connection.
  • Plus (+) icon: Adds a new OpenAI-compatible endpoint.
  • Eye icon: Reveals or hides the masked API key.
  • Gear icon: Opens settings to edit the endpoint URL or update the API key.

Section 2: Ollama API

This section is dedicated to connecting Ollama instances, commonly used for running open-source LLMs locally.

Purpose

  • Manage Ollama endpoints used for model discovery and chat requests.

Configuration fields

  • Endpoint URL: The address of the Ollama server (for example, http://192.168.18.39:11434).

Note: Ollama connections typically do not require an API key in this view.

Controls

  • Toggle switch: Enables or disables the Ollama connection.
  • Plus (+) icon: Adds a new Ollama endpoint.
  • Gear icon: Edits the endpoint details.

How to add a new connection

  1. Locate the section you want to add to:
    • OpenAI format API or
    • Ollama API
  2. Click the Plus (+) icon on the far right of the section header.
  3. Enter the Endpoint URL:
    • Example (Ollama): http://localhost:11434
    • Example (OpenAI-compatible): http://your-server:8000/v1
  4. If connecting to an OpenAI format API, enter the API Key.
  5. Turn the Toggle switch on to activate the connection.

Privacy and security notes

  • API keys are displayed as masked by default and should only be revealed when necessary.
  • Connections and credentials are stored within your IntraLLM AI instance according to your deployment’s configuration and access controls.
  • Limit Admin access to trusted users, since Admins can manage connection endpoints and credentials.