This n8n workflow creates an intelligent Telegram bot that interacts with users through a set of well-defined steps. When a user sends a message, the bot first identifies if it’s a command like `/start` or `/clear`, or a regular message. If `/start`, it responds with a welcome message explaining its features. If `/clear`, it deletes the current conversation history for that user and confirms the action. For regular text messages, the workflow manages chat history by storing the last 10 messages to maintain context and avoid token limits. It then formats the message and sends it to NVIDIA’s mistralai/mistral-nemotron API, requesting a context-aware AI response. The response is saved back into chat history and sent to the user, allowing for seamless ongoing conversations. The workflow also handles non-text messages by notifying users that only text is supported. This setup is ideal for creating a smart, memory-based customer support or personal assistant chatbot on Telegram, leveraging AI for meaningful, contextual interactions.