Prerequisites for setting up AI providers for Loop Genie

When you want to configure an AI provider such as OpenAI, IBM watsonx, Gemini, Claude or Ollama with IBM® DevOps Loop, ensure that the prerequisites are met.

Requirements for OpenAI integration

Ensure the following steps are completed before integrating OpenAI with DevOps Loop:
  1. Create an OpenAI account: Sign up on OpenAI’s platform and obtain API access.
  2. Generate an API key: Navigate to the OpenAI dashboard and create a secret API key for authentication.
  3. Install OpenAI library: Use Python or another programming language to install the OpenAI package.

Requirements for AZURE OpenAI integration

Ensure the following steps are completed before integrating AZURE OpenAI with DevOps Loop:
  1. Create a Model Deployment
    1. Sign In: Log into the Azure Portal.
    2. Open Resource: Click on your Azure OpenAI resource from your dashboard.
    3. Launch Studio: Select Go to Azure AI Foundry portal (or Azure OpenAI Studio).
    4. Go to Deployments: Click Deployments in the left navigation menu.
    5. Start Deployment: Select + Deploy model, then click Deploy base model.
    6. Choose Model: Select your preferred model (for example, gpt-4o) and click Confirm.
    7. Name & Save: Type a Deployment name, click Deploy, and save the name for your setup.

Requirements for Claude Desktop integration

Ensure the following steps are completed before integrating Claude Desktop with DevOps Loop:
  1. Create an Claude account: Sign up on Claude’s platform and obtain API access.
  2. Generate an API key: Navigate to the Claude dashboard and create a secret API key for authentication.
  3. Install Claude library: Use Python or another programming language to install the Claude Desktop AI package.

Requirements for Gemini integration

Ensure the following steps are completed before integrating Gemini with DevOps Loop:
  1. Create an Gemini account: Sign up on Gemini’s platform and obtain API access.
  2. Generate an API key: Navigate to the Gemini dashboard and create a secret API key for authentication.
  3. Install Gemini library: Use Python or another programming language to install the Gemini package.

Requirements for IBM watsonx integration

Ensure that you have the following details:
  1. API Key
  2. Project ID
  3. Endpoint URL

Requirements for Ollama integration

Ensure the following steps are completed before integrating Ollama with DevOps Loop:
  1. Install Ollama: Download the installer and install it.

    You must have read and understood the system requirements for each model of Ollama. for more information, refer to https://ollama.com.

  2. Download a model: Download the required large language model (LLM) such as Llama 2 or Mistral.
  3. Set up Python environment: Create and activate a virtual environment and install dependencies.
    Note:
    If you are using the Ollama instance that is bundled with DevOps Loop, you must follow these steps:
    1. Enable the instance in the Helm chart by setting the value to true in the values.yaml file:
      llama: 
    enabled: true
    2. Specify the required large language model (LLM) name in the Helm chart by providing the values in the values.yaml file:
      llama: 
    ollama: 
        models: 
        pull: 
            - <model name> 
            - <model name> 
        run: 
            - <model name> 
            - <model name>
      Where, the pull command is used to download a model, and the run command is used to run the model.