Build and launch a custom chat model in one flow. The setup moves a JSONL training file from Google Drive into OpenAI, starts a fine tuning job, and exposes a chat endpoint powered by the fine tuned model. Great for teams that need a focused assistant for one topic or brand voice.

On a manual run, the Google Drive node downloads your JSONL file as binary data. n8n then uploads the file to OpenAI storage with purpose set to fine tune and sends a POST request to create a training job on the base model. A separate chat trigger waits for messages and routes them to an OpenAI Chat Model that points to the fine tuned model id. An AI Agent bridges the chat and the model so people can talk to it in real time.

Set the Drive file id, your OpenAI API key, and the Authorization header for the job request. Expect setup time to drop from hours to minutes and responses to follow your training data closely. Use this for customer help, onboarding bots, and internal knowledge assistants where consistent tone and facts matter.

• Manual trigger to control when training jobs start
• Google Drive download of the JSONL training file as binary data
• OpenAI file upload with purpose set to fine tune
• HTTP POST to create a fine tuning job on the selected base model
• Chat trigger that listens for incoming messages via a webhook
• OpenAI Chat Model configured to use the fine tuned model id
• AI Agent that passes user messages to the model and returns replies

• Reduce model setup time from 3 hours to 10 minutes
• Automate over 90 percent of fine tuning steps end to end
• Cut upload and format mistakes by up to 80 percent
• Connect Google Drive and OpenAI in a single run
• Scale to new datasets by swapping one file id

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Google Drive and OpenAI. See the Tools Required section above for links to create accounts with these services.
3. In the n8n credentials manager, create a Google Drive OAuth credential. Name it clearly, then connect your Google account and grant file read access.
4. In the n8n credentials manager, create an OpenAI API credential. Get your API key from the OpenAI website under your API keys page and paste it into n8n.
5. For the Create Fine Tuning Job HTTP node, create a Generic Header Auth credential. Add an Authorization header with value Bearer YOUR_OPENAI_KEY and keep Content Type as application/json.
6. Open the Google Drive node and paste the file id of your JSONL training file. Make sure the binary property name is data.jsonl so the next node can find the file.
7. Open the Upload File node and confirm resource is file and purpose is fine tune. Set binary property name to data.jsonl so it picks up the downloaded file.
8. Open the Create Fine Tuning Job node and confirm the URL is https://api.openai.com/v1/fine_tuning/jobs and method is POST. Check the JSON body uses the uploaded file id and the base model you want.
9. Run the manual trigger to test. Check the Upload File node output for the file id and the HTTP node response for a job id and status.
10. After the fine tuned model is ready, copy the model id from OpenAI. Open the OpenAI Chat Model node and set the model field to that id.
11. Open the Chat Trigger node and copy the webhook URL. Send a test message to that URL or use the built in test view to confirm you get a response from the fine tuned model.
12. If you see file not found errors, verify the Google Drive file id and that your connected Google account has access. If OpenAI rejects the request, confirm the JSONL format is valid and your API key has fine tuning access.
13. For best results, keep each JSONL line valid and under token limits. Start with a small dataset, confirm the flow, then point the Google Drive node to larger files when ready.