Get Started

set VOLUME_PATH option in the .env file !!! set DEFAULT_USERNAME and DEFAULT_PASSWORD in the .env file point to the directory where you want to store large files. Models and knowledge bases can be more than 10 gb.

if you are on windows, in wsl2 you will need to set the memory in %UserProfile%\.wslconfig like

[wsl2]
memory=16GB

run with docker-compose, -d for detached mode, --build to rebuild the images

docker-compose up

after compose

navigate to the WebUI

all of your important settings are in Settings > Admin Settings > Models and Documents knowledge base and tools are under Workspace > Knowledge Base and Tools

choose a model like phi4 or llama3.2:3b

set extraction engine to http://tika:9998

maybe also specify

Chunk Size: 2000
Chunk Overlap: 200

you can download and then specify an embedding model like nomic-embed-text for uploading documents.

you can create a knowledge base and upload documents to it:

you can add a tool like ocrtool.py to the workspace

Custom models

Custom models allow you to configure system prompts, knowledge, and tools.

You can manage the user group a model belongs to

Pages for all the services

WebUI localhost:3000

Data Handler localhost:5000

Opensearch localhost:5601

Ollama localhost:11434

Opensearch localhost:9200

Tika localhost:9998

todo

automate as much openwebui setup as possible :/

ollama | llama_init_from_model: n_ctx_per_seq (2048) < n_ctx_train (131072) -- the full capacity of the model will not be utilized