Ingineer

Effortlessly plan events with friends

Comment

Inspiration

Making plans with friends should be easy. Instead, it often turns into a long group-chat loop:

“Are you free this weekend?” “Maybe.” “Where should we go?” “I don’t know.” “Someone pick something.”

The plan usually dies before anyone decides on a time, place, ride, reservation, or food.

Ingenium changes that by letting personal agents coordinate for people. Instead of waiting for every friend to reply, Ing can communicate with each person’s Ing to find when they are free. The group gets a real time to meet, then the right agents help turn that time into an actual plan. All designed around the app people already use to make plans: iMessage

What it does

Ingenium turns a group idea into a scheduled plan.

Text Ing something like: “Find a time for dinner this week,” “Plan Maya’s birthday,” or “What should we do Saturday?”

Ing can

  • Ask your friends’ Ings when they are free. Each person’s personal agent responds with availability, so the group does not need to wait for every person to manually check and reply.
  • Find the best shared time. Ing compares everyone’s availability and proposes a time the group can confirm.
  • Create the event. Once a timeslot is found, each person's Ing adds the event to their calendar
  • Call in specialists to make the plan real.
  • Find food options. A DoorDash specialist can return delivery options, estimated delivery times, group pricing, and direct links.
  • Plan rides. An Uber specialist can compare ride types, estimate costs, and create a link with the destination ready.
  • Find venues, reservations, tickets, flights, gifts, or supplies. Each specialist returns a few relevant options with the details needed to decide.
  • Build the final itinerary. After the group picks its options, Ing summarizes the plan: what is happening, when it is happening, where to go, and what each person needs to do.
  • Research anything that needs a browser. Users can ask Ing to compare restaurants, research events, or look up options. Ing works in its own cloud browser and sends screenshots and results back through iMessage.

How we built it

  • iMessage coordinator with Photon: We built Ing as the main agent users text. It understands the request, decides what help is needed, starts the group chat, and combines agent responses into one final plan.
  • Specialist-agent system: Ing can call in specialist agents for food delivery, rides, venues, restaurant reservations, tickets, shopping, flights, and itinerary planning. Each agent has a focused job and only responds when it is relevant to the conversation.
  • Real-time group coordination: Ing waits for the agents it called in to respond, then turns their results into a clear group plan instead of leaving users with scattered recommendations.
  • Calendar scheduling: We connected Google Calendar so Ingenium can check free/busy availability, find shared time slots in 30-minute intervals, and create an event after the group confirms.
  • Cloud computer: We gave Ing a sandboxed browser through E2B Desktop and OpenAI computer use. When a task needs deeper research, Ing can browse, compare options, send screenshots while it works, and return a short answer with links.
  • Privacy boundaries: All data from group chats are encrypted and only used by Ing. Calendar details are not shared with the group. Personal connected tools are only available in a user’s private conversation with Ing.
  • Web experience: We built Ingenium’s landing page with Next.js and Tailwind CSS.

Challenges we ran into

  • A group chat with multiple AI agents can quickly become overwhelming if every agent replies to every message. We addressed that by giving each specialist a narrow role and a strict reply rule: agents only answer when they are directly relevant, named, or responding to their own earlier message. Otherwise, they return PASS and stay quiet.
  • Ensuring user's data is safe, making sure what is sent in messages is encrypted and only used by Ing
  • How do we learn an individual's boundaries and preferences. ie. are they vegetarian? do they want to go out at 8am on a Monday just because their calendar is open?

Accomplishments that we're proud of

  • Built a working iMessage flow where a user texts one coordinator and that coordinator can add other agents to the group chat
  • Made specialist agents behave like focused collaborators rather than a pile of generic chatbots.
  • Created dynamic roles for food delivery, rides, venues, reservations, tickets, shopping, flights, and planning.
  • Built calendar availability matching that finds shared free time without revealing private event details.
  • Added event creation and invitation sending after the group confirms a plan.

What we learned

  • the hard part of multi-agent systems is not only getting agents to produce answers. It is deciding who should answer, when they should stay quiet, and how to turn several partial answers into one decision a group can act on.
  • trust matters more than having an agent do everything. Users should be able to see where recommendations came from, inspect links themselves, and keep the final decision for anything involving money, reservations, or personal information.
  • we learned that “social good” does not always need to look like a massive platform. Helping people overcome the small coordination barriers that keep them from seeing friends can be meaningful on its own.

What's next for Ingineer

  • Improve the planner so it can account for travel time, group size, dietary needs, budgets, and event timing in one final itinerary.
  • Test Ingenium with real friend groups to measure whether it actually reduces back-and-forth and leads to more completed plans.
  • Move from prototype-scale agent seats to a more reliable production architecture that does not depend on teammates’ laptops staying online.

Built With

Share this project:

Updates

Submission history