Use your local ChatGPT sign-in
The OAuth credential stays on this computer until you approve its encrypted SSH upload to your VPS.
Looking for your local Codex credential.
Credential updated (local time).
Connect to your VPS
Copy the address exactly from Hostinger. Your password is used only for this live SSH connection and is never saved.
Choose the detected n8n
Discovery uses read-only Docker commands. It does not execute anything inside n8n.
Review before installing
No existing n8n files or containers will be changed.
The wizard will
- Create or update
/docker/n8n-openai-oauth - Build and start only the
openai-oauthsidecar - Attach the sidecar to Not selected
-
Use the private hostname
n8n-openai-oauth - Upload the OAuth file with owner-only permissions
The wizard will not
- Edit or rebuild the n8n image
- Stop, restart, or recreate n8n
- Publish port
10531 - Create a Traefik route
The private bridge is ready
Copy the OpenAI credential values first, then use one of the node recipes below.
1. OpenAI credential
- Base URL
-
http://n8n-openai-oauth:10531/v1 - API key
-
local-only - Organization ID
- Leave empty
- Custom header
- Off
2. OpenAI Chat Model for AI Agent or Basic LLM Chain
- Model ID
-
Not detected - Responses API
- On for Chat Model node version 1.3
All models detected: Not detected
If the Responses API switch is absent, keep the node's default Chat Completions behavior; the bridge supports that route too.
3. HTTP Request node
- Method
POST- URL
-
http://n8n-openai-oauth:10531/v1/chat/completions - Authentication
- Generic Credential Type
- Generic Auth Type
- Bearer Auth
- Credential
openai-oauth- Bearer token
local-only- Authorization header
-
Bearer local-only - Content-Type
application/json- Send body
- On · JSON · Using JSON
- JSON body
-
The README includes complete AI Agent, Basic LLM Chain, JSON, and importable cURL examples. Start with no built-in tools. You can now return to Terminal and press Control+C to close this local wizard; the VPS sidecar will keep running.