Dantrons

Ask first

• Which shared calendar service does your family use, and how do you access its API?
• What messaging platform should receive conflict warnings?

Needs

• A family calendar with API access
• A messaging gateway (e.g. WhatsApp or Telegram)
• A persistent server or scheduler running a 10-minute cron

Build

1. Create a poller that fetches calendar events for the next 14 days every 10 minutes.
2. Compare start and end timestamps across all family members to detect overlaps.
3. If a new conflict is detected, format a warning message with suggested alternatives.
4. Send the alert to the family group chat and cache the conflict ID to prevent repeated notifications.

Check

• Add two overlapping events on the calendar and verify an alert arrives within 10 minutes.
• Confirm deleting one event clears the alert without duplicate warnings.

In my version

• Built in Node.js using the Skylight Calendar cloud API, Baileys WhatsApp WebSocket, and an in-memory cache running under PM2 on an Ubuntu VM.

Ask first

• What time should the morning and evening briefs be sent?
• What sources should be included (calendar, chores, weather)?

Needs

• Calendar integration
• Chore tracking system
• Messaging client
• Cron scheduler running in your local timezone

Build

1. Schedule a cron job for 7:00 AM and 7:30 PM in your local timezone.
2. At 7:00 AM, query today's calendar events, assigned household chores, and the daily weather forecast.
3. Synthesize the details into a concise, scannable group message.
4. At 7:30 PM, scan tomorrow's schedule to highlight required uniforms, musical instruments, or early departures.
5. Deliver each brief directly into the family group chat.

Check

• Trigger the digest endpoint manually and verify formatting, chore assignments, and event timings.

In my version

• Runs on node-cron in Australia/Sydney timezone, querying Skylight chores and calendar, and dispatching via Baileys WhatsApp gateway.

Ask first

• Where are your bank transactions stored (e.g. Supabase, spreadsheet, Plaid)?
• What spending categories and monthly limits do you want to monitor?

Needs

• Automated bank transaction feed
• Database or ledger of category budget limits
• Messaging endpoint for alerts

Build

1. Set up a twice-daily scheduled job to sync recent settled bank transactions.
2. Calculate month-to-date spending per category against configured monthly allowances.
3. Check for any category crossing 80% or 100% of its budget cap.
4. For uncategorized transactions, send a prompt asking for 1-tap categorization.
5. Send notifications with remaining days and daily burn rate allowance.

Check

• Insert a test transaction crossing the 80% mark and verify an immediate alert is sent.

In my version

• Uses Supabase database populated via Tiller Google Sheets bank feeds, evaluated by a custom TypeScript watchdog running on Google Cloud VM.

Ask first

• Where is your list of family birthdays and anniversaries stored?
• How many days of advance notice do you need for gift ordering and card delivery?

Needs

• Structured list of family milestone dates and relationships
• Daily morning cron job
• Group messaging connection

Build

1. Store family birthdays and anniversary dates in a cloud document or database.
2. Run a daily morning check comparing current date with upcoming milestones.
3. Trigger a 14-day advance alert to allow time for online gift ordering or shipping.
4. Trigger a 3-day reminder for final event logistics and card preparation.
5. Dispatch proactive reminders to the family messaging thread.

Check

• Set a test birthday 14 days from today and verify the concierge alert triggers at 9:00 AM.

In my version

• Runs on Google Cloud VM querying Firestore family memory collection, dispatching via Baileys WhatsApp.

Ask first

• What key financial metrics (savings rate, sinking funds, category variances) do you want displayed?
• What date and time should the monthly report generate?

Needs

• A PDF generation library (such as pdfkit)
• Access to your historical financial ledger
• Document attachment capability in your messaging client

Build

1. Schedule a cron job for 8:00 AM on the 1st day of every month.
2. Query the prior month's total income, fixed expenses, variable categories, and savings contributions.
3. Layout an executive document using vector tables, progress bars, and category breakdown charts.
4. Stream the generated PDF buffer directly to WhatsApp as a downloadable document attachment.

Check

• Run the PDF generator script locally and verify visual layout, typography, math accuracy, and WhatsApp delivery.

In my version

• Built with pdfkit in TypeScript, pulling from Supabase, rendered and delivered via Baileys sendDocument on Google Cloud VM.

Ask first

• Which supermarkets do you shop at?
• What are your strict family brand preferences and dietary substitution rules?

Needs

• Digital grocery list API
• Supermarket price lookup or catalog scraper
• Rule engine for preferred brands

Build

1. Intercept incoming grocery list items from voice notes, chat, or photo receipts.
2. Run each item through a rule engine to match exact preferred brands and package sizes.
3. Fetch real-time prices across target supermarkets for all items in the basket.
4. Calculate total cost for Store A, Store B, and an optimal split basket.
5. Apply a convenience threshold: if splitting saves less than $10, recommend the cheapest single store.

Check

• Add a staple item to the list and verify the bot returns matched brand alternatives and price comparisons.

In my version

• Built with a deterministic brand rule engine, Coles and Woolworths scrapers, and Skylight grocery list sync.

Ask first

• What inbound email address or webhook provider will receive forwarded messages?
• What calendar and budgeting tools should receive the extracted records?

Needs

• Inbound email webhook handler (e.g. Cloudflare Email Routing or Postmark)
• Multimodal LLM to parse text and PDF attachments
• APIs for your family calendar and accounting ledger

Build

1. Set up a dedicated inbound email address that triggers an HTTP webhook.
2. Parse the email MIME body, plain text, and any attached PDFs or images.
3. Pass content to a language model instructed to extract event dates, locations, and invoice amounts.
4. Call calendar API to create discovered events, and call budget API to record pending liabilities.
5. Send a confirmation message to WhatsApp summarizing what was added.

Check

• Forward a school newsletter PDF and verify all dates appear on the family calendar within 30 seconds.

In my version

• AtomicMail webhook receiver running on Express server, processed with Gemini multimodal structured output, syncing to Skylight and Supabase.

Ask first

• What audio formats or messaging channels will you speak into?
• What household actions should voice commands trigger?

Needs

• Audio transcription / speech-to-text model
• Streaming voice pipeline or voice note receiver
• Tool-calling agent capable of modifying calendar, lists, and chores

Build

1. Listen for incoming WhatsApp audio/ptt voice notes or browser WebSocket voice streams.
2. Convert OGG/Opus audio bytes into text using an audio-capable AI model.
3. Pass transcribed text into the agent's tool-calling loop to execute requested actions.
4. Generate a voice note or concise text reply confirming the action completed.

Check

• Send a 5-second voice note saying 'Add milk to the grocery list' and verify item is appended.

In my version

• Integrated into Baileys WhatsApp gateway and web voice endpoints, parsed by Gemini multi-step tool calls on Google Cloud VM.