n8n AI Terms

n8n · agreement dated 6 July 2026 · Official agreement

Read the agreement

2 of 16 cards in Use & Processing

Use & Processing · Company practice

With "bring your own LLM" features, the Customer picks the third-party AI models

For "bring your own LLM" AI Features, the Customer decides which third-party AI models connect to n8n's Services. n8n provides the workflow platform, and the Customer sets up the LLM providers using the Customer's own API credentials.

Supporting quotations

When using “bring your own LLM” AI Features, Customer controls which third-party AI models to connect to our Services. n8n provides the workflow platform and Customer configures LLM providers using Customer's own API credentials.

Qualifications & exceptions

  • Applies only to "bring your own LLM" AI Features.

Not stated here

  • The document does not list which AI Features are "bring your own LLM" features.

Use & Processing · Permission or license

n8n may use Usage Data and Customer prompts to improve AI Services, if aggregated and de-identified

n8n may collect and use internally Usage Data and Customer prompts to improve AI Services. This is allowed only while the Usage Data and prompts stay in an aggregated and de-identified form. The same document says n8n will not use Customer Content or Output to train ML models, and it does not say how that promise relates to prompts.

Supporting quotations

n8n may collect and internally use Usage Data and Customer prompts to improve AI Services, provided such Usage Data and prompts remain in an aggregated and de-identified form.

Qualifications & exceptions

  • Allowed only while the data remains aggregated and de-identified.
  • The use is internal.

Not stated here

  • The document does not define Usage Data or "AI Services", or say whether improving AI Services includes model training.
  • The document does not say whether Customer prompts count as Customer Content.

Cross-references

  • contrast (model-proposed): n8n will not use Customer Content or Output to train machine-learning models — n8n promises not to train ML models on Customer Content or Output, yet may use Usage Data and Customer prompts to improve AI Services in aggregated, de-identified form. The document does not say whether prompts are Customer Content or whether improving AI Services includes training, so the two provisions pull in different directions without a stated resolution.