Inspiration

As a daily learner trying to stay updated with global news while practicing foreign languages, I wanted a simple morning routine: waking up and immediately reading a structured, bilingual intelligence brief directly on my phone via Outlook, which is my primary mobile inbox.

However, existing AI agent automation hit a frustrating roadblock. While Gemini could run scheduled research tasks, attempting to dispatch these summaries to external inboxes triggered mandatory interactive confirmation popups. To receive my morning brief, I had to boot up my PC, open the browser, and manually click "Confirm"—completely defeating the purpose of automation whenever I was on the go or simply forgot.

I built AutoBrief to solve this real-world friction. It is a universal, cloud-native headless execution engine designed to automate recurring daily routines with true 100% zero-human intervention.

What It Does

AutoBrief operates as a two-stage decoupled execution pipeline across Google Cloud and Google Workspace:

  • Stage 1: Customizable Intelligence Generation (Gemini Spark): Acts as an adaptable content driver. Triggered on a scheduled heartbeat, it dynamically retrieves and synthesizes multi-source information across custom domains (e.g., global politics, tech frontiers, language-learning digests). It automatically commits structured output into Google Drive using a strict timestamp protocol (YYYY-MM-DD).
  • Stage 2: Extensible Headless Delivery (Google Apps Script Engine): Wakes via time-driven triggers, locates the daily document, strips rigid margins, and compiles raw text into a mobile-responsive HTML card container.
  • Autonomous Dispatch Gateway: Leverages system-level Google Workspace privileges (GmailApp) to silently route and dispatch the formatted brief to external inboxes (like Outlook) without ever blocking on manual UI confirmation popups.

Universal Multi-Task Matrix

While demonstrated as a daily news and language digest, AutoBrief is payload-agnostic:

  • Schedule & Priority Triage: Analyzes calendar events and priority deadlines to generate morning action cards.
  • Academic Literature Feeds: Scans PubMed/arXiv for domain-specific breakthroughs with intact citation links.
  • Automated Status Reporting: Compiles progress logs into executive-ready HTML emails.

How I Built It

  • Google Gemini (Spark Engine): Dynamic multi-source synthesis, translation, and scheduled prompt execution.
  • Google Drive State Store: Decoupled asynchronous state bus managing versioning protocols and buffering latency.
  • Google Apps Script (V8 Engine): Headless transformation layer injecting responsive CSS and preserving interactive hyperlinks.
  • GmailApp Gateway: System-level silent delivery channel that bypasses browser authorization prompts.

Challenges I Ran Into

  • Bypassing the Interaction Wall: Designing an asynchronous state relay via Google Drive so synthesis and delivery run independently without human-in-the-loop stalls.
  • Universal Mobile Rendering: Normalizing raw Google Docs exports into clean, mobile-responsive HTML cards that maintain hyperlink integrity across various email clients (Outlook, iOS Mail, Gmail).
  • Date Versioning Resilience: Standardizing string lookup protocols to ensure reliable daily state retrieval across timezones.

Accomplishments That I'm Proud Of

  • True 100% Autonomous Execution: The pipeline runs entirely in the cloud—no morning PC log-ins, no manual approval clicks, zero missed days.
  • Pristine Reading Experience: High-fidelity, mobile-first card layout that makes morning reading effortless.
  • A Reusable Foundation: Created a generic, plug-and-play execution engine capable of hosting any recurring task payload.

What's Next for AutoBrief

  • Multi-Channel Dispatch: Expanding delivery beyond email to Slack, Discord, and Telegram webhooks.
  • Bidirectional Interaction: Allowing users to reply directly to the daily email to dynamically adjust intelligence parameters or language difficulty levels.

Built With

Share this project:

Updates

Submission history