Support additional OpenAI-compatible providers for Lighthouse AI (e.g., DigitalOcean, GitHub)

[marticardus]

New feature motivation

Currently, Lighthouse AI in Prowler is tied directly to the official OpenAI API.
However, multiple cloud providers (such as DigitalOcean and GitHub) now offer OpenAI models through their own platforms.

Supporting these alternative endpoints would:

• Increase flexibility
• Reduce vendor lock-in
• Potentially lower costs depending on the provider chosen

Solution Proposed

Introduce a configurable option to specify:

• The API base URL
• Provider credentials

This would allow users to seamlessly switch between OpenAI’s official API and other providers that expose the same model interfaces (e.g., DigitalOcean, GitHub).

Implementation suggestion:

• Use an environment variable or configuration file to define the provider endpoint.
• Maintain full backward compatibility with the current default (official OpenAI).

Describe alternatives you've considered

1. Forking the codebase to hardcode alternative API endpoints – not practical, as it would make upgrades and maintenance difficult.
2. Proxying requests through a middleware service that rewrites calls from Lighthouse AI to the desired provider – adds unnecessary complexity and infrastructure overhead.
3. Sticking only with OpenAI official API – limits flexibility and ignores the growing ecosystem of compatible providers.

Additional context

No response

[andoniaf]

Hi @marticardus 🖖 thanks for the suggestion!

We're already internally exploring support for multiple providers, exactly for the reasons you described. I'll pass this feedback to the team and I hope we can share more details soon.

[andoniaf]

We added it to the public roadmap so you can upvote it and follow up the status: https://roadmap.prowler.com/p/support-multiple-llm-providers