Run a personal assistant through Telegram with eight agents
The author built a multi-agent n8n system with eight specialized agents controlled via Telegram by voice or text, replacing their personal assistant. Task: Suggested implementation: Build an n8n workflow featuring a GPT-4.1 orchestrator with PostgreSQL memory and eight specialist agents for email, calendar, expenses, travel, weather/news, research, content, social planning, and contacts. Users interact via Telegram text or voice, where the orchestrator routes requests to the appropriate agent to execute tasks using connected tools. Human involvement is required for initial setup, API key management, and monitoring agent outputs for accuracy.; tools: n8n. Human role: Source evidence is unavailable regarding specific human roles or manual interventions required for this workflow.; limitations: Prerequisites include an n8n instance, Telegram bot, Google Workspace access, PostgreSQL, and multiple API keys, creating significant setup complexity and ongoing maintenance costs. The source is a Reddit post with HTTP 403 errors, limiting verification of technical details and actual performance. Human approval is critical for sensitive actions like expense logging or social planning to prevent unauthorized operations..
Tools
Published on Agent Work Atlas:
Overview
The author built a multi-agent n8n system with eight specialized agents controlled via Telegram by voice or text, replacing their personal assistant.
Workflow
Suggested implementation: Build an n8n workflow featuring a GPT-4.1 orchestrator with PostgreSQL memory and eight specialist agents for email, calendar, expenses, travel, weather/news, research, content, social planning, and contacts. Users interact via Telegram text or voice, where the orchestrator routes requests to the appropriate agent to execute tasks using connected tools. Human involvement is required for initial setup, API key management, and monitoring agent outputs for accuracy.
Reported result
The author claims the system saves over 20 hours weekly, reduces travel planning time from hours to seconds, and achieves inbox zero while automating routine logging and research. These metrics are self-reported and lack independent verification or detailed measurement methodology. No evidence of system failures or specific error rates is provided in the source.
Human role
Source evidence is unavailable regarding specific human roles or manual interventions required for this workflow.
Limitations
Prerequisites include an n8n instance, Telegram bot, Google Workspace access, PostgreSQL, and multiple API keys, creating significant setup complexity and ongoing maintenance costs. The source is a Reddit post with HTTP 403 errors, limiting verification of technical details and actual performance. Human approval is critical for sensitive actions like expense logging or social planning to prevent unauthorized operations.
Try it
I want to pilot a multi-agent personal assistant using n8n and Telegram. First, set up an n8n instance and create a Telegram bot. [replaceable: provide your n8n URL and Telegram bot token]. Configure a core orchestrator agent using GPT-4.1 with PostgreSQL for memory. [replaceable: provide PostgreSQL connection details]. Start by connecting only the Calendar specialist agent. [replaceable: provide Google Calendar API credentials]. Send a test request via Telegram: 'Check my schedule for tomorrow.' Verify the orchestrator correctly routes this to the Calendar agent and returns the expected output. Measure the time taken compared to manual checking. If successful, proceed to add the Email agent. [replaceable: provide Gmail API credentials]. Do not enable expense or social agents until human review confirms data handling is secure. Ask clarifying questions if API scopes are unclear.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.