n8n Workflow: Triage Russian Support Requests to JSON

Triage Russian support requests into structured JSON—automatically

This n8n workflow takes a Russian support request, sends it to TokenTool’s OpenAI-compatible chat completions API, and returns a clean JSON triage with category, priority, and a draft reply. Perfect for faster routing and consistent responses in your support queue.

What this workflow does

Designed to run on demand, it demonstrates the full triage pipeline end-to-end:

• Runs manually to start a synthetic demo request.
• Creates a sample Russian support request and sets the TokenTool text model ID to use.
• Sends the request text to TokenTool’s endpoint /api/v1/chat/completions via an HTTP request.
• Uses a system prompt that instructs the model to output JSON containing: category, priority, and reply.
• Returns the model response, with the generated JSON available in choices[0].message.content.

Setup (TokenTool + model)

• Create a TokenTool API key and add it to an n8n HTTP credential using Authorization: Bearer ....
• Replace REPLACE_WITH_CURRENT_MODEL_ID with a current TokenTool text model ID.
• Run the workflow manually and confirm the output JSON format before using real requests.

Use cases

• Route Russian-language customer messages to the right support team by category and priority.
• Generate a draft reply to speed up agent response time while keeping tone consistent.
• Test triage logic before connecting a real intake channel (e.g., webhook/form) and mapping output fields to your helpdesk workflow.

Technical details

• Manual Trigger for on-demand execution
• HTTP Request to TokenTool /api/v1/chat/completions (Bearer authentication)
• Set node for request/model variables
• Includes UI/support nodes like Sticky Note

Requirements: TokenTool account with a positive balance, an active API key, and a current text model ID. For full setup guidance and a public sanitized workflow, see: https://tokentool.ru/instructions/n8n.