---
id: hackathon-event-details
title: "MCP Apps Hackathon — Event Details, Rules, Prizes"
type: research
status: active
owner: ivan
created: 2026-02-19
updated: 2026-02-19
tags: [hackathon, research, event, prizes, rules]
---
# MCP Apps Hackathon — Event Details, Rules, Prizes
## Overview
**MCP Apps Hackathon (aka ChatGPT / MCP Apps Hackathon)** — Feb 21, 2026 at Y Combinator HQ, San Francisco.
Hosted by **Manufact** (formerly mcp-use, YC S25; co-founders Luigi Pederzani/@pederzh & Pietro Zullo/@pietrozullo). Focus: Build/deploy working MCP (Model Context Protocol) apps/servers with UIs that plug directly into ChatGPT, Claude, Cursor, VS Code, etc. Use their open-source mcp-use SDK (~9k GitHub stars, 5M+ downloads), Inspector for testing/preview, and deploy to Manufact MCP Cloud (GitHub connect).
**Official pages:**
- https://manufact.com/hackathon (apply/registration)
- https://events.ycombinator.com/manufact-hackathon26 (YC listing + agenda)
## Schedule (exact from YC events page)
- **9:00 AM** — Check-in & Team Formation
- **10:00 AM** — Kickoff & Workshops
- **10:30 AM** — Hacking begins
- **12:00 PM** — Lunch
- **6:00 PM** — Demos & Judging
- **7:00 PM** — Awards & Wrap-up
Full-day in-person event at YC office (SF). Food/drinks provided. Hacking window ~7.5 hours.
## Prize Structure & Breakdown
Public details are high-level (no full tiered breakdown published; organizers emphasize "the winner"/"best team").
- **Grand Prize (1st / Best Overall)**: Guaranteed YC interview + **$10K OpenAI credits** + swag + additional prizes from supporters. Recent posts upgraded to "Mac Minis (last ones in SF)" + "tens of thousands of $$ in credits" total.
- **Other prizes**: Swag + "prizes from our supporters" (likely smaller credits/swag for runners-up or category winners). No explicit 2nd/3rd cash tiers or separate "YC interview track" announced.
- **Sponsors**: Anthropic, Cloudflare, OpenAI, **Puzzle** (puzzlefin — co-promoting heavily), WorkOS. Expect sponsor-themed awards (e.g., "Best use of Puzzle API", best Anthropic/Cloudflare integration), but **no specific categories or prize values published**. People from OpenAI/Anthropic are judging.
**YC interview goes to 1st place / best team only** (phrasing across all announcements: "the winner secures...", "best team gets...").
## Rules & What Counts as "Pre-Built"
No official rules doc/FAQ published (common for small in-person events). Based on all public info:
- **In-person only**, build during the official hacking window.
- **Heavily encouraged/expected**: Start with mcp-use SDK starters/templates, Inspector, and deploy to Manufact Cloud. No prohibition on using official templates, examples, or boilerplates (this is the point of the event).
- **Pre-seeded data / external repos**: Not addressed, but with only ~7.5 hrs, realistic projects will use seeds/starters. No "must be 100% from scratch" rule mentioned.
- Standard hackathon norms apply (no copying entire prior projects). Projects must be MCP-based and demo-able live.
- Deploy via GitHub → Manufact Cloud for instant availability in AI chats/agents.
## Team Rules
- **Solo allowed** (explicit team formation slot at 9 AM; Reddit threads show people seeking teammates).
- **Max team size**: Not specified (typical 4 max for in-person YC events; form on-site).
## Presentation Format
- **Live demos at 6 PM** (core of judging). Short pitches + functional demo expected (MCP app running in ChatGPT/Claude/etc.). No mention of slides/video-only; in-person live format.
## Participants
- Limited by YC office space + "request a spot" (hosts confirm they'll reach out to confirm).
- Recently "few spots left" → "3x the number of applicants" → "biggest MCP Apps Hackathon." Likely 50-150 people (not a 500-person mega-event).
## Judging Criteria
**No published rubric or weights.**
Judges include Jon Xu (YC General Partner) + reps from OpenAI/Anthropic.
Inferred from event focus + similar MCP hackathons:
- Innovation / creative use of MCP
- Technical quality & execution (SDK integration, deploy, UI/functionality)
- Demo quality / real-world usefulness in AI chats/agents
- Sponsor tool usage (bonus for Anthropic/Cloudflare/Puzzle/WorkOS)
All decisions final; live demos central.
## Key Takeaways
- **YC interview** → Winner only (1st/best team).
- **Sponsor prizes** → Likely exist (Puzzle especially prominent) but not detailed publicly.
- **Rules loose on starters/templates** → Use official mcp-use resources heavily. Pre-seeded data fine for demo.
- **Very short build time** → Focus on quick, impressive MCP integration + live demo.
Sources: manufact.com/hackathon, events.ycombinator.com, X/LinkedIn announcements, Reddit threads.
