Pryvet icon

Pryvet

Privacy-first AI chat completions via Pryvet

Chat → Completion

AI-generated

Summary

Send a chat completion request to Pryvet with a sequence of role-tagged messages, a selected or custom model, a privacy mode, and an optional web search setting. The node requires at least one message with a role and non-empty content, then returns Pryvet’s response as JSON output items.

Inputs

  • Messages (required) — Add one or more messages, each with a User, Assistant, or System role and non-empty text content.
  • Model (required) — Choose a listed model (default: GPT-5.6 Terra), or select Custom and provide its API identifier in Custom Model.
  • Custom Model — Required when Model is Custom; enter the exact model identifier.
  • Mode — Privacy mode for pseudonymization: Open sends data as-is, Normal applies standard pseudonymization, or Strict enables additional recognizers. Defaults to Normal.
  • Web Search — Enable web search for the model. Defaults to off.

Output shape

One or more JSON output items containing Pryvet's response for each input item.

Each request uses Pryvet's chat completions endpoint and includes the configured project ID. With Continue On Fail enabled, an item that errors returns an object with an error field; otherwise errors are thrown. 401 and 402 responses are reported as invalid API key and payment required errors, respectively.

Links

Discussion