How to build local AI with GPT-OSS, Ollama and n8n
Step1: Download and setting Docker (docker.com). You might need to restart your computer to finish setting up Docker.
Step 2: Setting n8n local
Run script in CMD "docker volume create n8n_data" to create volume n8n_data.
You can run "docker volume ls" to check whether it has been created.
After that, run script:
docker run -it --rm --name n8n -p 5678:5678 -e GENERIC_TIMEZONE="Asia/Ho_Chi_Minh" -e TZ="Asia/Ho_Chi_Minh" -e N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS=true -e N8N_RUNNERS_ENABLED=true -v n8n_data:/home/node/.n8n n8nio/n8n
Note: You can remove "--rm" if you want to keep this container. And now, your script is:
docker run -it --name n8n -p 5678:5678 -e GENERIC_TIMEZONE="Asia/Ho_Chi_Minh" -e TZ="Asia/Ho_Chi_Minh" -e N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS=true -e N8N_RUNNERS_ENABLED=true -v n8n_data:/home/node/.n8n n8nio/n8n
You can check your time zone here List of tz database time zones - Wikipedia
After running that script, you will see the following output and can go to
n8n.io - Workflow Automation:
Step 3: Install Ollama
Step 4: Install gpt-oss model
You can use this script "ollama run gpt-oss:latest"
Note: you might set system variable for Ollama and use Windows Power Shell instead of CMD.
Step 5: Use Ollama chat model inside n8n
Run script "ollama serve" to turn on ollama service.
You can create your workflow in n8n following this image:
Detail each node:
- LLM Chain:
- Edit Fields:
URL: http://host.docker.internal:11434 => Save
After that, chosse model: gpt-oss:latest
Finally: You can click button "Execute workflow" to test your workflow or use Chat box in the left conner.
Nhận xét
Đăng nhận xét