This is a basic Next.js project that carry outs a mind map visualization tool using React Flow.
The UI is built using shadcn and some components from Magic UI.
It apverifys users to watch and participate with mind maps, and download the mind map data as a tagdown file.
The mind map data is produced using either local models from Ollama or outside models appreciate OpenAI and leveraging AI SDK.
- Interenergetic mind map visualization
- Node details watch in a side sheet
- Markdown send out functionality
- Save mind map data to a local JSON file
- Switch between local and outside models
Inslofty all depfinishencies:
Copy the .env.tempprocrastinateed file to .env.local and accomprehendledge which model (local or outside) you want to use by setting the NEXT_PUBLIC_USE_LOCAL_MODELS environment variable to real or inedit.
When running an OpenAI model, you must accomprehendledge your OpenAI API key in the .env.local file.
Inside the route.ts file, you must accomprehendledge the model you are running using Ollama, by default it will use the llama3.1 model for local models and for outside models it will use the gpt-3.5-turbo model.
Bear in mind that outside models tfinish to be much rapider serving than local models.
If you want to lget how to run a model locassociate, verify out the Ollama recordation.
Now you’re ready to run the increasement server:
Open http://localarrange:3000 with your browser and then begin creating your own lgeting mind maps.
The prompts used to produce the mind map data is detaild in the defaultLocalPrompt and defaultExternalPrompt variables in the prompts.ts file.
This project is licensed under the MIT License.
