Why We're Building the Aggregation Layer for Predictions

Prediction markets are exploding, but the experience is painfully fragmented leading to an experience no better than traditional sportsbooks.

Today we're launching Agg.Market, a prediction market aggregator built to give users the best prices, deepest liquidity, and widest selection across every prediction market from a single interface.

We've been heads down building this for months. Here's why we think aggregation is the winning model, and how we're approaching it differently.

Why Prediction Markets

Prediction markets are one of the most exciting categories in finance right now. The 2024 election cycle proved that mainstream demand exists at scale, and the design space is still wide open. Volume has surged >400% Y/Y with many analysts projecting $1T+ in volume by 2030.

But the current experience is fragmented. Liquidity is siloed across platforms. Users are overpaying on fees, leading to similar outcomes offered by traditional sportsbooks. Onboarding still feels like 2021 DeFi. Today there's no easy way to compare prices or find the best market for a given question.

While prediction markets are a controversial topic today, an increasingly polarized media landscape makes a decentralized source of truth necessary. We’re building the foundational source of truth for predictions, and saving users an average of >3% per trade in the process.

The Space is Crowded, Been There, Done That

We've competed in crowded categories for years, and we've learned that being early doesn't matter as much as being right on the size of market + opportunity. The best product (and GTM) wins, and we see prediction markets as the venture-scale opportunity we've been looking for.

Aggregators Win

The core thesis behind Agg.Market is straightforward: in competitive marketplaces, selection wins.

The platform that offers the deepest liquidity layer, most market selection, and best fees will compete directly with giants Kalshi + Polymarket. This playbook has played out in every major marketplace category.

I wrote about this years ago while still at DoorDash — building a marketplace with the best selection wins. At DoorDash we won by building a selection flywheel and onboarding the most restaurants. Amazon won early by offering every book title and then expanding rapidly to new categories. The best market selection + deepest liquidity will win in predictions.

**Agg.Market** is Kayak for prediction markets. Just like Kayak lets you search every airline and hotel from one place and find the best price, we let you search every prediction market and find the best odds.

We’re building the best aggregation technology for the category, think Relay or 1inch applied to predictions enabling smart routing + order splitting to optimize every trade. One interface, every market, best execution.

Agg doesn’t just aggregate @kalshi and @polymarket, we offer all of the leading venues with support for @trylimitless, @predictdotfun. With @myriadmarkets, @opinionlabsxyz, and @hyperliquidx via HIP-4 ready as fast follows.

The Distribution Thesis

Here's the part that gets us most excited: More than half of prediction market volume will eventually flow through front ends that aren't prediction market startups. Robinhood, Coinbase, and DraftKings are already live today. ESPN, Uniswap, and other major consumer front ends won’t be far behind.

These platforms have massive existing user bases and strong distribution, but they don't want to build prediction market infrastructure from scratch. They need an aggregation and execution layer to support them.

While we have a consumer-facing product, we’re building a predictions stack for B2B. We already have 5 customers committed and dozens of conversations in progress with platforms exploring prediction market integration.Initial build partners include DeepBot, Telegram’s first integrated wallet with >700K users. In conversation, the pattern is consistent: these companies want to offer predictions but don't want to build matching engines, liquidity aggregation, and multi-platform execution from the ground up.

That's Agg.

How We're Building It

We're not just slapping a UI on top of multiple APIs. The technical approach matters, and we've invested heavily in the infrastructure layer.

Market Matching: We use a multi-LLM approach with in-house reinforcement learning to match equivalent markets across platforms. When the same question is listed on Polymarket, Kalshi, and elsewhere with slightly different wording, our system identifies and links them — so users always see the full picture of available liquidity.

Smart Routing + Order Splitting: Once markets are matched, our execution engine handles optimized routing. Large orders get split across platforms for best execution, similar to how DEX aggregators route swaps across liquidity pools.

Best-in-Class UX: We've abstracted away the complexity. Users login with any Web2 auth or non-custodial wallet and account creation is serverless via @portal_hq, no seed phrases or browser extensions. Order execution is optimized via @relayprotocol. The outcome is one-click trading at the best available price, regardless of which platform is filling the order.

We Want Your Feedback

We're building this in public because we think the best products are shaped by the people who use them. What features do you want to see? How can we make the experience better?

My DMs are open! Please reach out directly if you have ideas on how we can improve Agg.Market or are interested in building on Agg.