Create an AI assistant using n8n that will:
Receive incoming messages from clients via WhatsApp (various languages: Kazakh, Azerbaijani, English, etc.).
Automatically detect the client's language and translate the message into Russian.
Send this message to me via Telegram with the original, translation, and language indicated.
I respond in Russian, and the AI:
translates my reply back into the client's language,
sends the translated reply to the client via WhatsApp.
🔧 Tools:
n8n (installed on my server with public access)
WhatsApp API (preferably via Twilio or 360dialog — specify which is more convenient)
Telegram Bot API — for forwarding messages to me and receiving my responses
OpenAI API (ChatGPT) — for translating text and detecting language
📚 Functional requirements:
1. Incoming messages from the client (WhatsApp)
Webhook in n8n receives the incoming message.
Language of the text is determined automatically via GPT or another API.
Message is translated into Russian (using OpenAI or DeepL).
Translated text + original + language are sent to Telegram.
2. My reply (via Telegram)
I reply directly in Telegram (bot).
Reply is sent back to n8n.
Message is translated from Russian into the client's original language.
Reply is sent to the client via WhatsApp.
📩 Example flows:
Example 1:
Client (in Kazakh) → WhatsApp → translation into Russian → I receive via Telegram:
💬 New WhatsApp client:
Language: Kazakh
Original: [client's text]
Translation: [into Russian]
Example 2:
My reply in Telegram (in Russian):
Good afternoon! We deliver to your region.
AI translates into Kazakh and sends to the client via WhatsApp.
⚠️ Additional:
It is necessary to save the client's language (by phone number) to avoid detecting it each time.
It is desirable to keep a chat log (can be in Google Sheets, Airtable, or a text file).
Ability to adapt the system in the future for other messengers (Instagram, Telegram, Viber).
Kyiv