Cursor × Thrad
London2026
London · 2026 · 28 May

Two tracks. One conversational stack.

Ad infrastructure for AI — ship agents that operate the same real-time prompt, intent, and placement surface brands and publishers use in chat-native inventory.

Build in Cursor; tension-test buy-side, sell-side, and measurement workflows the way a production conversational stack would — fast, contextual, and explicit about policy boundaries.

10max score
7judges
5live demos
3podium places
2tracks
Why this moment

Ads moved into the chat box. Agents move the infrastructure.

Thrad frames it plainly: ad infrastructure for AI — publishers monetize with in-chat placements; brands buy impressions in LLM-native channels. That stack monitors prompts, scores intent, ranks creative, and deploys sponsored answers in milliseconds. This hackathon is the agentic layer on top: who automates bids, who guards brand safety, who proves ROI back to the conversation — with humans clear on what ships autonomously.

In February 2026 OpenAI flipped the switch on ads inside ChatGPT — CPC bids of $3–$5, $200K minimum advertiser commitments, Expedia and Best Buy live on day one. Anthropic answered with Super Bowl spots saying “ads are coming to AI, but not to Claude.” Thrad co-hosts — same real-time bidding layer powering paid ads in LLMs, used by Fortune 500 brands and case-studied by AWS and BCG. You are building in Cursor — the strongest agentic IDE available — and the ask is concrete: ship agents that buy, sell, or measure attention inside conversations, and make it obvious what they will and will not do.

@sualehasif996

Scaling Shopify, turbopuffer, and databases

A conversation with @sirupsen on scaling Shopify through flash sales, collaborating across infra teams in the 2010s, and what is next for databases.

@kamil_sattar

Shopify AI Toolkit for your favorite coding agent

Shopify AI Toolkit connects Claude Code, Cursor, VS Code, Gemini CLI, Codex, and more to Shopify docs, schemas, validation, and your store.

@Shopify

The Shopify AI Toolkit is here

Manage your store with Claude Code, Codex, Cursor, VS Code, and more — connected to first-party Shopify tooling.

@glenngabe

OpenAI launches an ads signup landing page

New landing page includes sign-up flow, ads documentation, and a walkthrough video; OpenAI verifies each advertiser before they go live.

@samboboev

Agentic commerce may start with builders, not baskets

a16z partner @nlevine19 argues early agentic commerce could center developers and solopreneurs accessing tools, APIs, and datasets — before everyday consumer purchases.

@JeremyMearsX

Buyers can DM merchants inside Agentic Commerce chats

No detached contact forms — shoppers ask questions in-thread, agents relay them, and merchants answer without breaking the conversational surface.

Track 01 · Advertiser side Buy-Side Agents

Buy-Side Agents

Agents that plan, bid, and buy ads inside LLM channels — ChatGPT-style placements, Thrad-powered apps, sponsored conversational answers. The risk is direct: misplaced spend, off-brand creative, or runaway budget. The human-in-the-loop question: when can the agent commit money alone, what counts as a brand-safety violation, and who signs off on creative before it serves?

  • Auto-bids on conversational intent, escalates spend spikes
  • Generates and rotates ad creative inside brand guardrails
  • Detects wasted spend, pauses unprofitable placements

Build for the advertiser: brands, agencies, DSPs.

Track 02 · Publisher side Sell-Side & Measurement

Sell-Side & Measurement

Agents that help AI publishers monetize and prove ROI — picking placements, scoring prompt-level intent, attributing conversions back to chat sessions. The risk is indirect: a bad signal becomes a bad billing event. The human-in-the-loop question: when is intent strong enough to monetize, and which conversions need a human-confirmed audit?

  • Scores prompt intent in real time, gates ad eligibility
  • Stitches attribution from chat → click → conversion
  • Flags suspicious traffic, asks before refunding spend

Build for the publisher / SSP / measurement side: AI apps, chatbots, attribution platforms.

Bonus prizes

Four sponsor-aligned buckets — judges award up to one point each (see rubric).

  • Tavily

    Best use of Tavily

    Search and research API built for agents — grounding, retrieval, and live data in your hack workflow.

    Tavily → Docs →
  • Overmind

    Best use of Overmind

    Ship the clearest integration or workflow that shows Overmind doing real work in the product — not a throwaway badge.

    Overmind → Docs →

  • Best use of Cursor

    Standout use of Composer 2.5 and the Cursor SDK — agentic coding in the IDE plus programmatic agents, tools, and patterns from the examples repo.

    Composer 2.5 → SDK docs →
  • Alpic

    Best use of Alpic

    Host, deploy, or route MCP servers and agent tools with Alpic — show a real integration judges can trace from your repo to a live endpoint.

    Alpic → Docs →
Panel

Judges

Get to know them and what they like. They are calling the shots on winners this evening.

Awards

Prizes

Five tiers. Top three walk out with Cursor Ultra, Thrad credits, and swag — fourth and fifth take home Thrad credits.