Run email, calendar and tasks from a Telegram voice assistant
How it works:This project teaches you to create a personal AI assistant named Jackie that operates through Telegram. Task: Suggested implementation: Build an n8n workflow triggered by Telegram messages. The system detects voice or text input, transcribes voice via OpenAI, and routes text to an AI assistant (Jackie) using OpenRouter. Jackie accesses Gmail, Google Calendar, and Google Tasks via APIs to summarize emails, check schedules, or manage tasks. A human must review and approve any outbound messages, bookings, or changes before execution.; tools: n8n. Human role: The original post does not say how a person stayed involved.; limitations: Requires valid API keys for Telegram, OpenAI, OpenRouter, and Google OAuth2. Integration complexity is high due to multiple service dependencies. Voice transcription accuracy depends on audio quality. The source does not specify costs, latency, or failure handling for API rate limits. Human approval is critical to prevent unauthorized actions..
Tools
Published on Agent Work Atlas:
Overview
How it works:This project teaches you to create a personal AI assistant named Jackie that operates through Telegram.
Workflow
Suggested implementation: Build an n8n workflow triggered by Telegram messages. The system detects voice or text input, transcribes voice via OpenAI, and routes text to an AI assistant (Jackie) using OpenRouter. Jackie accesses Gmail, Google Calendar, and Google Tasks via APIs to summarize emails, check schedules, or manage tasks. A human must review and approve any outbound messages, bookings, or changes before execution.
Reported result
No measured outcomes or success metrics are reported in the source. The workflow is described as a template for personal productivity, but actual performance, error rates, or user satisfaction are unknown. A measurable acceptance check would be: 'Verify that 100% of voice messages are correctly transcribed and that email summaries accurately reflect unread inbox contents in a test environment.'
Human role
The original post does not say how a person stayed involved.
Limitations
Requires valid API keys for Telegram, OpenAI, OpenRouter, and Google OAuth2. Integration complexity is high due to multiple service dependencies. Voice transcription accuracy depends on audio quality. The source does not specify costs, latency, or failure handling for API rate limits. Human approval is critical to prevent unauthorized actions.
Try it
Act as an n8n workflow builder. I want to create a personal AI assistant named Jackie on Telegram. Inputs: [Telegram Bot Token], [OpenAI API Key], [OpenRouter API Key], [Google OAuth2 Credentials], [Target Gmail/Calendar/Task IDs]. Steps: 1. Set up a Telegram Trigger node to listen for messages. 2. Add a branch to detect if the message is voice or text. 3. For voice, use OpenAI Whisper API to transcribe. 4. Pass the text to an AI Agent node (using OpenRouter) with tools for Gmail (read unread), Calendar (check events), and Tasks (create/list). 5. Format the AI response in Markdown. 6. Send the response back via Telegram. Safeguards: - Do not enable write actions (sending emails, creating events) without a manual approval step in the workflow. - First, run a dry test with a sample voice message and a text query 'What are my unread emails?'. - Show me the JSON output of the transcription and the AI's tool calls before finalizing. - If any API key is missing, stop and ask for it. Do not guess credentials.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.