I cant use any llm except google i dont know what to do. I really want to use olllama and i have done what not. I have fintuned context number i have redownloaded the ollama llm, ive restarted bolt multiple times, i have reinstalled multiple times. And i was still fine but now gemini crashes because my project has been huge ig idk. Please help
Hi @sff10147,
bolt cannot handle big projects at the moment, cause its sending almost everything every time to the provider instead of just the needed stuff.
So if your project is to big to work with bolt.diy, I would recommend to use another tool for now, like Windsurf/Cursor/Aider.
Thank You. Could you help me with ollama not working too. It does not even recognise a model present. I have already tried LM Studio too. And when i try to switch models in a big file, it does not work but if i use the same model in a fresh project it starts to work. And only Google, Mistral and Hugging Face work right now.
Share come screenshots of your configuration please and how you setup ollama.
=> The model/provider is not capable of managing the project size in this case
=> Ollama/LM Studio, you could try to set a higher context size and see if this helps (depends on your model and graphics cards / system, if you can do that)
i think my system is not capable. Is there an online method to edit such big files? I am really stuck.
I used deepseek chat do to all the things manually too but i exceeded the limit i am so stuck
could you tell me what are the minimum requirements of the model
Did you try to use Google Gemini 2.0? It´s free and works well normally.
I´ve build my task-list-advanced nearly fully with it.
Alternative: If bolt also cant handle it anymore with Gemini, you can switch to desktop tools like Cursor/Windsurf/Aider to finish your project.
It´s a workaround, but sometimes needed at the moment. Hope in the near future bolt can handle everything more efficient and also does DIFF etc., so it gets much faster and more performant.