Independent community-led hackathon · OpenAI support requested

Local problems.
Working agents.

A focused build day where 100 selected makers work with Indore companies, NGOs and civic organizations to turn real challenges into usable AI products.

01 / Start with a verified problem 02 / Build with responsible data 03 / Finish with a working link

This is not a pitch-deck competition.

Every team starts with a scoped local problem and finishes with something a real person can open, test and question.

Every submission includes

  1. 01A working product link
  2. 02A public source repository
  3. 03A 90-second product demo
  4. 04Setup notes and known limitations
  5. 05A responsible-use and data note

Track 01

Make repetitive operations easier to run.

Useful internal search, support triage, multilingual customer help, document workflows and small-team knowledge systems.

Data rulePublic, anonymized or synthetic information only.

Problem-partner outreach is in progress. Organizations will be named only after confirmation.

Bring one problem.
Meet twenty possible approaches.

Companies, NGOs and civic teams can submit a narrow challenge that matters to their users. We help turn it into a safe, buildable brief before it reaches participants.

1

Submit

Describe the user, current process and desired improvement.

2

Scope

We remove sensitive data and define a realistic success measure.

3

Mentor

A domain representative answers questions during the build sprint.

4

Review

Useful winners enter a 30-day pilot-discovery clinic—without guaranteed procurement.

Submit a challenge brief

A day designed around protected build time.

Check-in and setup clinic

Accounts, tools, starter resources and team readiness.

Opening and responsible-use briefing

Rules, judging, data boundaries and what counts as shipped.

Problem partner briefings

Real users, real constraints and open questions.

Team formation and build sprint

Groups of four or five begin with mentor office hours.

Midpoint product review

Working-path check: cut scope, test the core and document decisions.

Submission freeze

Live link, repository, demo and limitations are locked.

Demos, judging and awards

Shortlisted teams show the product—not slides about the product.

25%

Working implementation

The core path runs reliably during the live demo.

20%

Thoughtful AI use

Codex or the OpenAI API creates material product value.

15%

UX and accessibility

A first-time user can understand and operate it.

15%

Safety and reproducibility

Limitations are honest and another builder can run it.

20–25working submissions
100%documented limitations
30 daysfollow-up for finalist teams
1 reportpublic outcomes and learnings

Hosted by The Origin Guild

Built on repeated event operations—not first-event optimism.

The team has supported 42+ community meetups, workshops and hackathons across AI, developer tools and emerging technology.

97GitHub Copilot Dev Days attendees
114Monad Blitz builders
600+largest conference attendance
Is OpenAI officially sponsoring this event?

Support has been requested and is not yet confirmed. The event is independently organized by The Origin Guild. This page will be updated transparently if support is approved.

Who can apply?

Developers, students, founders, designers and domain experts who can commit to building and presenting a working product. Selection will prioritize readiness, diversity of skills and fit with the problem tracks.

Do I need an existing team?

No. Individual applicants can join structured team formation after the problem briefings.

Will partners share private user data?

No. Challenge briefs must use public, anonymized or synthetic information. Teams may not collect or expose sensitive personal data during the event.

Does winning guarantee a contract or government adoption?

No. Selected projects receive a follow-up clinic and may explore a pilot with the relevant partner. Procurement, funding or deployment is never promised in advance.

Indore · 24 October 2026

Bring the problem.
Leave with proof.