Oscar Levy and Antonin Parrot were running into the same irritation every trader hates: toggling between screens, managing multiple logins, hunting for the best price across platforms that should, by all rights, talk to each other.
Only they weren't trading stocks or bonds. They were working prediction markets—Kalshi and Polymarket, specifically—and discovering that as hedge funds crowded into event contracts, the infrastructure hadn't kept pace. Fragmented liquidity. Clunky APIs. No unified view of positions. The kind of friction that, in more mature asset classes, disappeared years ago.
So they built something. River Markets, a prime brokerage for prediction markets, went live in late May with backing from Y Combinator's Spring 2026 batch. The pitch: a desktop terminal and unified API that stitches Kalshi and Polymarket into a single interface, routing orders to wherever liquidity is best and claiming an average of two cents per contract in price improvement.
According to a Y Combinator LinkedIn post celebrating the launch, River has already moved north of $50 million in volume and counts three of the top ten traders on both venues among its early users. These metrics are reported by the company and lack third-party verification, and the company—still just Levy and Parrot, according to Y Combinator's directory as of late May—hasn't disclosed funding details beyond standard YC terms.
Still, the timing is striking.
Infrastructure Rush
River's arrival coincides with a broader institutional land grab in prediction markets, a space that spent years as a curiosity for political junkies and crypto enthusiasts before Wall Street decided it might be worth taking seriously.
Interactive Brokers rolled out a unified prediction markets interface on May 14, aggregating Kalshi, CME Group's ForecastEx, and IBKR's proprietary ForecastEx into a single screen with automatic best-price routing—part of a broader wave of similar efforts by other firms. For IBKR clients, that meant no more juggling separate venue accounts—the exact problem River is solving, just for a different audience.
Then there's Clear Street. The prime brokerage and clearing firm became Kalshi's first institutional FCM member on May 1, offering hedge funds block trade capabilities and ETF swap solutions tied to event contracts. That partnership, along with similar moves by Marex, got extensive coverage in finance industry publications earlier this year, part of a narrative that accelerated after the CFTC dropped its March 12 Advance Notice of Proposed Rulemaking on event contracts.
River positions itself somewhere in the middle of this ecosystem: more sophisticated than consumer aggregators, but lighter and nimbler than the IBKR or Clear Street behemoths. Whether a two-person startup can compete with those incumbents is an open question. But perhaps the market is big enough—or fragmented enough—that multiple routing layers can coexist.
What River Actually Does
The core mechanics are straightforward, if not simple. Instead of logging into Kalshi and Polymarket separately, traders route orders through River's system, which evaluates liquidity in real time and splits fills across venues when advantageous. Server-side execution algorithms handle the rest: icebergs, pegs, take-profit and stop-loss orders—table stakes for equity traders, but relatively novel in prediction markets.
River standardizes contract identifiers across venues using what it calls "river_ids," a normalization layer that lets traders reference the same event across different exchanges without wrestling with venue-specific ticker symbols. The order management system consolidates positions and profit-and-loss across venues. Subaccounts allow for segregated capital management within a single River account, useful for funds running multiple strategies or managing client money.
According to the company's documentation, live connectivity covers Kalshi, Polymarket, and Polymarket US. Novig and Rothera are listed as in integration, though some discrepancies exist between the marketing page and technical docs, which were crawled on different dates in late May—a reminder that early-stage startups often move faster than their own websites can keep up.
A Python SDK hit version 0.3.0 on May 22, maintained by a developer identified as "oscarlv-rm" on PyPI. The SDK offers Ed25519-signed REST and WebSocket APIs, asynchronous order book streams, and programmatic access to River's routing engine. For systematic traders, that's the real product: not the terminal, but the pipes underneath.
Levy, a co-author on two arXiv papers published in February on lead-lag trading and forecasting in prediction markets, appears to be both the academic and the operator. It's a profile that's become familiar in the YC ecosystem: researcher-turned-founder, building tools that started as internal workflows and grew into commercial products.
Client List, With Caveats
River lists Chimera Capital Management, Game Point Capital, Cleat Street, and Skywalk as clients on its homepage. These client relationships are listed by River Markets but lack independent verification, and the company has not disclosed contract sizes or usage details. The claim to power three of the top ten traders on both Kalshi and Polymarket comes from the same Y Combinator social post that cited the $50 million volume figure—essentially, company-reported metrics without third-party confirmation.
That's not unusual for a startup that's been live for a matter of weeks. But it does leave some questions hanging.
Not the Only Play
River is hardly alone in attacking cross-venue fragmentation. Three other Y Combinator companies in the 2026 cohorts are building in adjacent territory, which suggests either a real market opportunity or a crowded field heading for consolidation. Probably both.
Valence, from the Winter 2026 batch, offers aggregation and arbitrage-surfacing tools with backtesting APIs aimed at systematic traders. Sequence Markets, also Winter 2026, focuses on cross-exchange execution across crypto, prediction, and tokenized-asset venues. Outside the YC universe, Snag Solutions launched agg.market on April 28—a consumer-facing aggregator with an internal central limit order book and seed funding led by Gradient.
River's bet is that professional and institutional traders represent a distinct segment, one that cares more about execution quality, server-side algos, and API-first design than retail participants do. The distinction matters, especially as the market bifurcates between consumer tools and infrastructure aimed at systematic desks managing larger position sizes.
Whether River can defend that positioning against Interactive Brokers, which already serves many of those same institutional clients, is unclear. IBKR has distribution, regulatory infrastructure, and a multi-decade head start in building trader trust. River has... speed, focus, and access to venues IBKR doesn't (yet) support natively, namely Polymarket.
That might be enough. For now.
Regulatory Tailwinds, Maybe
The CFTC's March rulemaking notice marked something of a policy shift, formalizing oversight of prediction markets after years of regulatory ambiguity. According to a KPMG advisory report published this spring, combined Kalshi and Polymarket volume exceeded $40 billion in 2025, with Kalshi's fee revenue estimated at $263.5 million as part of KPMG's advisory analysis. More than a dozen designated contract market applications have been filed since early 2025, reflecting industry-wide momentum toward regulated event contracts.
That regulatory clarity, assuming it holds, could accelerate institutional adoption—or slow it, depending on what the final rules look like. Either way, the arrival of routing layers like River, IBKR's unified screen, and Clear Street's FCM partnership suggests prediction markets are transitioning from niche retail product to something resembling a traded asset class with real Wall Street infrastructure.
For Levy and Parrot, the next challenge is scaling a two-person operation to compete with firms that have hundreds of engineers and decades of institutional relationships. The early metrics and Y Combinator backing give them a runway. Whether it's long enough remains to be seen.
