Tutorial

[ZachZuk]

Is it possible if I could get a more in depth tutorial on how to download and install, I am not experienced in this area, some help would be great.

[tsol]

In short:

1. git clone the repo, make npm install
2. Rename .env.default into .env
3. Register an application for search in your google account, get API key and Context ID from there
4. Fill in .env with those
5. Rename models_skel folder into models
6. download symantics model files into some other folder, using "git lfs install", "git lfs clone [repo]" - move all .json files and model.onnx into models/bla-bla.../default
7. download main GGML model using "node download-model.js" (9.5 GB)
8. git clone "llama.app" repository from github into llama.app folder inside of project
9. build llama.cpp (using make) - file main should be created

Run test file using "npm run start" - this will run "run.ts" file, which calls search tool with a query.

Then create your own tools in src/prompt-tools folder according to search.js format.

Sounds a little bit complicated, i guess some of it can be automated.

[ZachZuk]

Thanks, do I need to install anything like JS or is git all I need?

[tsol]

You will need node environment installed. All the rest pretty much is installed using "npm install" inside the repository after you've downloaded it. Have fun! :)