nilAIInstall

Summary

Sends each input item to a selected nilAI model for inference inside its TEE. Combines optional Instructions and required Input into the prompt, then returns the model response with optional signature and enclave attestation verification.

Inputs

• Model Name or ID — Select a model loaded from the configured nilAI endpoint, or provide a model ID using an expression.
• Instructions — Optional task directions. The node trims this text and places it before Input in the combined prompt.
• Input (required) — Content for the model to process. The node combines it with Instructions, separated by a blank line when both are present.
• Simplify Output — Defaults to on. Returns text, tee_verified, attestation, model, signature, and usage. Turn off to return all fields from the raw response, with tee_verified and attestation added.
• Verify TEE Signature — Defaults to on. Fetches the endpoint's public key and checks the response signature; the result is included as tee_verified. If disabled, tee_verified is null.
• Verify Enclave Attestation — Defaults to on. Checks the serving enclave's AMD SEV-SNP attestation and includes the result in attestation. This check runs once before processing the input items.

Output shape

One output item per input item, paired with its corresponding input.

With simplified output, text is the extracted output text or null if no text is found; model, signature, and usage are null when absent. Attestation verification returns an object, including an error and attestation_verified=false if verification fails. Errors stop processing unless the node is configured to continue on fail, in which case an error item is returned for that input.

Links

• n8n community nodes installation guide