How to use Gemini API key with GitDiagram?

[pradyothpnv]

Hi @ahmedkhaleel2004,

Thanks for creating GitDiagram! I really like how it helps visualize complex codebases by generating Mermaid.js diagrams using OpenAI models.

I read in the README that most of the prompt engineering and processing happens in /backend/app/prompts.py, and currently, the project uses OpenAI’s o4-mini .

I want to switch to using the Gemini 2.5 Flash API key instead.

Could you please advise:

What changes are required to use the Gemini API key in this project?, Which file(s) or configuration should I update to replace GPT-3 or o4-mini with Gemini?, Any other considerations or recommended approach for integrating Gemini here?

Thanks in advance for your help!

[viralcodex]

Hi @ahmedkhaleel2004,

Thanks for creating GitDiagram! I really like how it helps visualize complex codebases by generating Mermaid.js diagrams using OpenAI models.

I read in the README that most of the prompt engineering and processing happens in /backend/app/prompts.py, and currently, the project uses OpenAI’s o4-mini .

I want to switch to using the Gemini 2.5 Flash API key instead.

Could you please advise:

What changes are required to use the Gemini API key in this project?, Which file(s) or configuration should I update to replace GPT-3 or o4-mini with Gemini?, Any other considerations or recommended approach for integrating Gemini here?

Thanks in advance for your help!

Hi @pradyothpnv,
If u want to use Gemini then you will have to host locally, and use your own API Key, for that you can clone this repository and checkout to my PR where I have added the Gemini API Support file. You can use that by adding you Gemini API Key in .env in root folder.

here are the commands to checkout into my PR where my changes are:
origin being this repository url you used to clone it and 116 is my PR number, you can see it from here #116

git fetch origin pull/116/head:pr-116
git checkout pr-116

Hope this helps!!