Hackathon 2026 · Concluded

AI Agents for Real Life

Five focused days in July. Solo builders and teams of up to three were asked for one thing: an AI agent that solves a real problem in a real community. Most had written their first prompt only weeks earlier. Winners were announced on July 16, 2026.

5Mission tracks
7Awards given
5Day build
6Judges
Solo or teams of 3Open to all cohorts and alumniPrize sponsor: Base44July 6 to 10, 2026
The projects

What our builders shipped

Five mission-track winners chosen by the judging panel, plus two Community Favorites voted by participants. Every project below was built in five days.

Track winner · Education & Career

CareerFolio

Atisha Banjare, Vaishali Goel, Hansneet Randhawa

Turning what you did into proof of what you can do. Transforms a resume into an evidence-backed professional portfolio, with role-matched career paths, recruiter-ready bullets, STAR stories, an elevator pitch, and a 3/6/12-month growth roadmap.

The problem. Professionals, students and career switchers struggle to translate their work into credible proof of skill. CareerFolio helps them build credibility, not just publish a portfolio.

Why it matters. It collapses the distance between doing great work and being able to prove it. Career coaching that scales to thousands at once, without fabricating a single number.

LovableChatGPTClaudeNatively AIGeminiAI/ML API
Track winner · Financial Empowerment

Budget Buddy

Tammy Chan

A budgeting and savings coach. Visualizes financial status, tracks progress toward goals, and encourages better spending habits without lecturing.

The problem. Financial planning isn’t part of most school curriculums. Reckless habits perpetuate debt and affect both mental and physical wellness.

Why it matters. Widespread adoption of a genuinely easy budgeting tool lifts financial wellness across demographics and reduces the societal cost of the financial planning gap.

Make.comGoogle GeminiGoogle FormsGoogle SheetsGmail
Track winner · Future of Work

Pepper

Mike Jones

An adaptive chief of staff for neurodivergent brains. Type out everything on your mind and Pepper returns one clear next step, with a short list for today and everything else quietly parked.

The problem. For many neurodivergent people the hard part isn’t doing the work, it’s starting. Most to-do apps show all forty items at once and expect the user to plan, which is the exact thing these brains find hardest.

Why it matters. Pepper takes prioritizing off your plate. It runs a crisis check on every input, uses language that never shames, and keeps personal data off the model.

Claude CodeBase44Multi-agent orchestration
Track winner · Health & Wellbeing

HealthSignal

Malar Manogar

Public health intelligence made simple. Turns trusted public health data into clear guidance, answering what’s happening, how serious it is, and what to do about it in under thirty seconds.

The problem. Public health information is widely available but hard to understand and act on. HealthSignal makes it current, trustworthy, location-specific and usable.

Why it matters. A transparent multi-agent architecture: deterministic agents verify freshness and detect trends, and the model only communicates findings that have already been verified. That reduces hallucination in a domain where accuracy isn’t optional.

LovableReactTypeScriptClaude Sonnet APIEPA AirNowDelphi EpidataCDC NNDSS
Track winner · Personal Productivity

Camp House of AI

Lisa Erickson · The One-Person AI Company

The self-running business. A fully autonomous lead-to-client engine disguised as a vintage lakeside summer camp. Visitors sign a guestbook, receive one AI skill daily, share wins in a Lodge, and are looked after by an autonomous AI Keeper while the founder does nothing daily.

The problem. Non-technical founders, especially women over 40, become the bottleneck in their own businesses. AI tools are fragmented and intimidating.

Why it matters. The founder fired herself as the bottleneck. Content self-publishes by date, the Keeper handles every guest, and gamified honors drive retention with zero daily founder effort. It proves autonomy by being autonomous.

Claude APIAgentic workflowsLovableVercelHand-built HTML/CSS/JSManus
Community Favorite · Solo

MumPreneur AI

Varsha Kamble

A marketplace for mum-built micro-apps. A platform where mothers on career break can list, sell and get paid for the AI tools and automations they have built for small businesses.

The problem. Millions leave the workforce for childcare and lose both income and career confidence. Even after gaining AI skills there’s no clear path to monetize them, while small businesses badly need affordable AI tools.

Why it matters. AI acts as the confidence layer. It writes each builder’s listing, prices the work so she never undersells, and drafts the professional replies she is nervous to write.

Claude APILovableStripe
Community Favorite · Team

CareLink

Shirin Shaikh, Amruta Ambre, Charu Arora

An AI voice companion for elderly parents. A voice-first companion helping seniors stay safe, healthy and independent, with medication reminders, everyday conversation, and emergency alerts to family caregivers.

The problem. Millions of families struggle to help elderly loved ones stay safe and independent from a distance. Existing solutions are expensive, complex, or cover only emergency response.

Why it matters. CareLink detects phrases like "I fell" or "I need help" and immediately alerts designated family members, while keeping caregivers informed with shared updates every day.

GPT-OSS 120BOllamaClaude Code
Mission tracks

Five problems worth solving

Builders picked one track that matched the problem they wanted to solve. Each track carried its own award.

Health & Wellbeing

Agents that support mental health, fitness, nutrition, sleep, or community care.

Mood check-in coach · Nutrition planner · Caregiver support

Financial Empowerment

Agents that help people make better financial decisions and improve economic opportunity.

Budgeting coach · Financial literacy assistant · Scholarship finder

Education & Career Growth

Experiences that help people learn, upskill, and reach their career goals.

Study companion · Resume and interview coach · Mentorship matching

Personal Productivity

Agents that help individuals study, organize, focus, learn, or get unstuck.

Study buddy · Inbox triage · Focus and habit tracker

Future of Work & Entrepreneurship

Agents that empower workers, founders, freelancers, and small businesses.

Customer research assistant · Pitch deck reviewer · Operations helper

How it ran

Apply, build, reveal

Applications closed June 30. Five focused build days, July 6 to 10. Winners announced directly on July 16, with no live demo day.

How projects were judged

Weighted on impact, not complexity

A panel of six industry judges scored every submission across five dimensions. Community Favorites were decided by participant voting after submissions closed.

30%

AI innovation and use of AI

Is AI central to the solution? Does it go beyond adding a chatbot? Is AI used thoughtfully to create value?

25%

Functionality and execution

Does the prototype work? Is the experience clear and usable? Does it demonstrate the intended functionality?

20%

Problem clarity

Is the problem clearly defined and worth solving? Does the team understand the people they’re building for?

15%

Impact potential

Could this positively affect individuals, organizations, or communities? Is there room to grow or scale?

10%

Presentation and demo

Is the solution communicated clearly? Can a non-technical audience understand the value being created?

Every entry

What we asked for

A title and track, a 300 to 500 word description, a working prototype, a two to three minute demo video, the AI stack used, and an impact statement.

This is not a competition to build the most complex technology. We are looking for solutions that solve real problems, demonstrate thoughtful use of AI, and create meaningful impact. A simple solution that addresses a genuine need often performs better than a complex one with limited real-world value.

Guidance given to every participant
0Countries reached
0Cities
0Community members
ScholarshipsNo one turned away
7Awards given

The next cohort starts January 4, 2027

AI Launchpad 2026 is complete. The January 2027 cohort runs ten weeks across ten professional tracks, and applications open soon.

See the January cohort