Dify

install

¹

git clone --depth=1 https://github.com/langgenius/dify
cd dify/docker
cp .env.example .env

usage

sudo docker compose --env-file .env up -d

Create chatbot and knowledge bases

²

Dify → <user> → Settings → Model Provider

Ollama → Setup

Model Name: llama3.1:8b
Base URL: http://<your_host>:11434

Ollama → Add Model

Model Type: Text Embedding
Model Name: mxbai-embed-large:latest
Base URL: http://<host>:11434

Model Type: Text Embedding
Model Name: jina/jina-embeddings-v2-base-en
Base URL: http://<host>:11434

OpenAI → Setup

API Key: <apikey>
API base: https://api.openai.com

OpenAI → Show Models
Dify → Studio → Chatbot → Create from Blank
```
APP icon & name: llama 3.1
```
Studio → llama 3.1 → <model> CHAT → llama3.1:8b
Dify → Knowledge → Create Knowledge → Upload files → Save & Process → Go to Documents
```
Index mode: Economical
```
Studio → llama 3.1 → Context → Add → <knowledge>

resource

cache

dify