Inception PlatformInstall

Summary

Sends a chat conversation to the Mercury 2 model and returns a generated completion.

Inputs

• Use JSON — Switch to true to supply the conversation as raw JSON in the Messages (JSON) field instead of using the structured message builder.
• Messages (required) — Add one or more messages with a role of system, user, or assistant and free-form content. This field is used when Use JSON is false. At least one message is required.
• Messages (JSON) (required) — When Use JSON is true, provide a JSON array of objects with role and content, e.g. [{"role":"user","content":"..."}]. This field is required in JSON mode.
• Reasoning Effort — Controls the model's reasoning effort. Allowed values: instant, low, medium, high. Defaults to medium.
• Additional Fields — Optional request parameters merged into the API body: max_tokens (number, min 1, default 8192), temperature (0–1, default 0.75), top_p (0–1, default 1), and stop (list of strings where generation should stop).

Output shape

The complete response object returned by the Inception Platform API.

One output item is produced for each input item. Each output item's json contains the raw API response from POST /chat/completions, including the model's reply, choices, and usage data. If continue-on-fail is enabled, a failed item is returned as { error: error.message }.

Examples

Example 1: A workflow that asks Mercury 2 a question and returns the model's answer.

Send a user message to Mercury 2 with medium reasoning effort.

Links

• n8n community nodes installation guide