Skip to main content

Configure an OpenAI-compatible endpoint

These settings apply after the privately distributed Local LLM extension is installed in your organization (see Introduction).

Settings in Azure DevOps​

After the Local LLM extension is installed and assigned to your project:

  1. Open Organization settings or Collection settings.
  2. Find ClearSpecs AI and open the Local LLM tab.
  3. Set:
    • Base URL - HTTP(S) endpoint implementing an OpenAI-compatible chat/completions API (trailing slashes are normalized).
    • Model - Model name your server exposes (for example an Ollama model tag).
    • API key - Optional; include if your proxy requires it.

Common default for Ollama’s OpenAI bridge: http://127.0.0.1:11434/v1.

Settings are stored in Azure DevOps Extension Data for the extension; access follows Microsoft’s scoping for extension data in your organization.

CORS (browser and Ollama)​

The Boards UI runs on an Azure DevOps origin (for example https://dev.azure.com). Calls to a local Ollama instance are cross-origin. The LLM server must return CORS headers that allow your Azure DevOps web origin.

For Ollama, set OLLAMA_ORIGINS (or your version’s documented CORS settings) to include your Azure DevOps origin (scheme + host). Without that, browser fetch requests fail even when Ollama is running locally.

Privacy​

This SKU does not call the ClearSpecs cloud assistant API for prompt execution; requests go only to the URL you configure (subject to your network and proxy rules).