Order groceries from health data and a meal plan
Austen Allred's AI automatically orders groceries based on body scans, health data, and preferences, mapped to a meal plan.
Agents people built themselves, lesser-known products, and ones the author didn't name.
Agents people built themselves, lesser-known products, and ones the author didn't name.
Austen Allred's AI automatically orders groceries based on body scans, health data, and preferences, mapped to a meal plan.
A clinic's automated phone AI handled the poster's appointment booking call instead of a human.
A commenter called their doctor about a sick pet and the AI booked an appointment 30 minutes later, which worked out.
An AI reviewed a contractor PO, emails and line items and drafted dispute emails, spotting issues the user missed and leading to $20K credited back.
An agent named Anna, reached by text or WhatsApp, manages family email, calendar and school WhatsApp groups and automates forms and schedules.
A commenter ingests social content with Obsei and posts across networks via getlate.dev with an AI agent, except Reddit.
Drew McAllister's agent flagged counteroffers by text, built owner follow-up lists, tracked bids, drafted emails, read lease redlines and flagged deadline conflicts.
An agent researches keywords, publishes one SEO article daily and refreshes articles every two weeks using Search Console and Ahrefs data.
A toolkit gives a coding agent a semantic layer and dashboard files so it can query data warehouses and publish in-depth reports and visualizations.
A commenter uses an AI agent connected to QuickBooks Online, email and Drive to check the books and turn supplier invoices into bills.
A CLI tool built on Twilio gives terminal agents a voice to make calls, with a dashboard for questions they can't answer.
An agent opened the TTD website for Tirupati, picked options, entered details and took the booking to the payment stage without human clicking.
Robin Delta built a voice agent that answers support calls around the clock, resolves tier-1 issues or escalates, and books appointments.
Feynman multi-agent system searches papers and the web, synthesizes findings, runs simulated peer review, verifies citations, and outputs a cited brief.
A user's agent tried to book a dentist appointment, and the dentist's agent countered, leading to negotiation over timing, fees and arbitration with American Express.
The poster built an agent that compares real delivered prices, tests every coupon it finds, reads owner feedback and goes through checkout without ordering.
The Midas Markets agent researches assets, analyzes portfolios and PnL, finds tokens by momentum, and executes swaps and perp positions on Hyperliquid.
An agent sent firm follow-ups every 48 hours, built a paper trail and prepped a regulator grievance; the $1,200 flight refund was processed within a week.