The goal of this package is to provide an AI assistant to the world of Pokémon.
It consists in a stack of services, listed in the docker-compose.yml file.
In a nutshell, it encompasses an UI and an inference service. A custom agentic proxy intercepts the requests between these services, processes them, and eventually augments them with information from a vector DB.
The models have been selected with respect to their minimalism, performance and multilingualism.
The project has been set-up such as French 
This project can also be seen as a natural language processing exercice with relatively limited resources, i.e. a gaming computer. The specs it has been built with are:
- a linux/amd64 platform ;
- git and docker ;
- 32 Go RAM ;
- a Nvidia GPU (12 Go VRAM) with cuda.
To make use of the later, the Nvidia container toolkit is needed.
Start by cloning the repo:
git clone https://github.com/almarch/pokedex.git
cd pokedexThe project uses nginx as a reverse proxy.
Server-client transactions are encrypted using SSL keys. For more security, use a domain name and a CA certificate.
Generate the SSL keys:
openssl req -x509 -nodes -days 365 -newkey rsa:2048 -keyout services/nginx/ssl/ssl.key -out services/nginx/ssl/ssl.crt -subj "/CN=localhost"Also, the UI service requires a secret key:
echo "WEBUI_SECRET_KEY=$(cat /dev/urandom | tr -dc 'A-Za-z0-9' | fold -w 32 | head -n 1)" > .envThe project is containerized with docker.
Pull, build & launch all services with compose :
docker compose pull
docker compose build
docker compose up -d
docker psOllama is included in the stack.
It requires 2 models:
- a LLM. By default, Mistral-Nemo is selected.
- an encoder. By default, BGE-M3 is selected.
If you change the models, adjust services/myAgent/myAgent/__main__.py accordingly.
Pull the models from the Ollama container. If Ollama runs on container 123:
docker exec -it 123 bash
ollama pull mistral-nemo:12b-instruct-2407-q8_0
ollama pull bge-m3:567m-fp16A Qdrant vector DB is included in the stack.
It must be filled using the Jupyter Notebook service, accessible at https://localhost:8888/lab/workspaces/auto-n/tree/pokemons.ipynb.
Pokémon data come from this repo.
Open-WebUI is included in the stack.
Reach https://localhost:8080 and parameterize the interface. Deactivate the encoder model, and make the LLM accessible to all users. If needed, make accounts to the family & friends you would like to share the app with.
This framework can readily adapt to other agentic projects.
- The data base should be filled with relevant collections.
- The custom agentic logics is centralised in
services/agent/MyAgent/MyAgent/MyAgent.py.
Say we need to tunnel the server using a VPS. In other terms, we want some services from the GPU server, let's call it A, to be accessible from anywhere, including from machine C. In the middle, B is the VPS used as a tunnel.
| Name | A | B | C |
|---|---|---|---|
| Description | GPU server | VPS | Client |
| Role | Host the services | Host the tunnel | Use the Pokédex |
| User | userA | userB | doesn't matter |
| IP | doesn't matter | 11.22.33.44 | doesn't matter |
The services we need are:
- The web UI and the notebook, available at ports 8080 and 8888 respectively.
- A SSH endpoint. Port 22 of the gaming machine (A) will be exposed through port 2222 of the VPS (B).
The ports are pushed to the VPS:
ssh -N -R 8080:localhost:8080 -R 8888:localhost:8888 -R 2222:localhost:22 userB@11.22.33.44The SSH ports 2222 and 8080 have to be opened.
sudo ufw allow 2222
sudo ufw allow 8080
sudo ufw reloadThe UI is now available world-wide at https://11.22.33.44:8080.
The jupyter notebook is pulled from the VPS:
ssh -N -L 8888:localhost:8888 userB@11.22.33.44The notebook is now available for the client at https://localhost:8888.
And the VPS is a direct tunnel to the gaming machine A:
ssh -p 2222 userA@11.22.33.44Note that userA, not userB, is required for authentication ; idem for the password.
This work is licensed under GPL-2.0.