Large Language Models

You can add your own LLM API keys to your Codio organization to use in your Course from the Organization > LLMs page in Codio.

LLMs Keys

We support the following providers:


Custom LLM Providers

You can add a custom LLM provider, such as Google Gemini, to Codio by entering the API keys on the Organization > LLMs page. This allows you to use your chosen LLM provider in your courses.

Custom LLM Providers

Steps to Add a Custom LLM Provider

  1. Click Add provider and enter the following details:

  • Name: A recognizable name for the provider.

  • Provider: The name of the LLM service.

  • Endpoint: The API endpoint URL for the provider.

  • API Key: The authentication key needed to access the provider.

  • Authentication: Select the authentication type (e.g., Header or Query Param).

  • Auth Template: Specify the authorization format.

  • Endpoint Environment Variable: Define an environment variable for the API endpoint.

  • API Key Environment Variable: Define an environment variable for the API key.

  1. Click Create to save and activate the provider.

Examples of how to use a custom LLM provider

OpenAI:

  • name: your openai custom name (custom value)

  • provider: openai (custom value)

  • endpoint: https://api.openai.com/v1

  • API key: your API key from openai

  • authentication: Header

  • auth template: Authorization: Bearer {{apikey}}

  • endpoint environment variable: OPENAI_CUSTOM_URL (custom value)

  • API key environment variable: OPENAI_CUSTOM_KEY (custom value)

After saving the parameters in the organization and enabling the provider in the course, run the script below in the assignment.

#!/bin/bash

SUB_PATH=/v1/chat/completions
LLM_URL=$OPENAI_CUSTOM_URL$SUB_PATH

curl -X POST $LLM_URL -H "Authorization: Bearer $OPENAI_CUSTOM_KEY" -H "Content-Type: application/json" -d '{"model": "gpt-3.5-turbo", "messages": [{"role": "system", "content": "your question here" }]}'

Gemini:

  • name: your gemini custom name (custom value)

  • provider: gemini (custom value)

  • endpoint: https://generativelanguage.googleapis.com

  • API key: your API key from gemini

  • authentication: Query param

  • auth template: key={{apikey}}

  • endpoint environment variable: GEMINI_CUSTOM_URL (custom value)

  • API key environment variable: GEMINI_CUSTOM_KEY (custom value)

The script:

#!/bin/bash

SUB_PATH=/v1beta/models/gemini-2.0-flash:generateContent
LLM_URL=$GEMINI_CUSTOM_URL$SUB_PATH?key=$GEMINI_CUSTOM_KEY

curl -X POST $LLM_URL  -H "Content-Type: application/json" -d '{"contents": [{"parts":[{"text": "your question here"}]}]}'

Note

SDKs may require or expect standard environment variables (e.g., OPENAI_BASE_URL, OPENAI_API_KEY) or explicit values, while curl can use the custom names you configure.

Enabling LLM for Courses

When keys are added, you can enable use of the key in a course. See Course LLMs.

To enable Codio LLM keys for your organization, please contact help@codio.com to initiate the process. Specify whether you require OpenAI or Anthropic keys.

LLM Organization Usage

LLMs Usage

Note

The names showing in this screenshot are example names.

Usage will be shown for:

  • Daily, weekly, monthly and annual usage

  • Daily and monthly top 10 consumers by course

  • Each course total usage

Note

The values shown are estimates and may not reflect the exact numbers.