AI Assistant · Messaging · Memory

Personal AI Assistant with Memory

Prototype

A Telegram-based AI assistant that keeps session memory and returns structured responses, with several integrations still planned.

Full n8n canvas for the personal AI assistant, showing the Telegram trigger, AI agent, tool connections, and vector store

Problem

A personal assistant needs to hold conversational context across messages and respond in a structured, useful format rather than one-off replies.

System Design

An n8n-hosted AI agent triggered over Telegram, with session lookup, a connected memory store, and an explicit ask-for-approval tool wired into its available actions.

Workflow Stages

01Telegram message
02Session lookup
03AI model
04Structured response
05Formatting
06Telegram reply
07Memory storage

Tools Used

n8n
Telegram
Supabase
Groq
AI models
MCP integrations

Safety & Human Review

Explicit human-approval step before certain assistant actions

Implementation Evidence

Real screenshots of the working system — collapsed by default so this page stays readable.

What I'd Improve Next

Planned Gmail, Calendar, Tasks, and document-memory integrations are not yet complete.

Have a similar workflow to automate?

I'll tell you honestly whether automation is the right fix.