Market making and arbitrage. Every layer in-house.
We build proprietary market-making and cross-venue arbitrage strategies, run for our own book across centralized exchanges, perp DEXs, and prediction markets. Every layer of the engine is ours, from market data and research to strategy, execution, and risk.
Two strategies. One engine.
Market making and arbitrage are the business, run in-house for our own book. Both run on our proprietary engine, and every layer of it is ours, from market data up to execution. So every quote and every spread is priced off data we captured, normalized, and can replay.
Two-sided quotes on order books, sized and skewed by inventory and risk limits. Run for our own book, on the venues and markets where depth is thin and spreads are wide.
- Quoting driven by the full order book, not top-of-book
- Inventory and risk limits enforced in the engine
- Proprietary, run in-house
Cross-venue price and funding dislocations, found on one clock and captured on both legs. Across centralized and decentralized venues, wherever the same asset trades at different prices.
- Cross-venue spreads and funding carry
- Both legs priced off the same normalized data
- Built for venues others don't connect
The foundation both strategies run on. Native feeds from each venue, normalized onto two schemas (perps and prediction markets) and captured on a single clock, replayable bit-for-bit.
- 13 venues targeted
- 3 venue classes (centralized, perp DEX, prediction)
- Deterministic, replayable capture
- Market data available on request
From native feed to filled order.
One proprietary Rust engine runs the whole path. The data that prices a quote or flags a spread is the same data we capture and replay, so what we test is what we trade.
Connect to each venue's native market-data feed, whether it's a centralized exchange, a perp DEX, or a prediction market.
Map every venue onto one schema per market type and one clock, so books and prices compare across venues.
Market-making and arbitrage logic reads the normalized state, including quotes, spreads, funding, inventory, and risk.
Orders go out through the venue's own API, inside hard risk limits. Every event is captured for replay and review.
We study live and captured market data to find where spreads open and how order books behave, then test strategy changes on replayed data.
13 venues targeted across centralized exchanges, perp DEXs, and prediction markets.
What brings you here?
Pick what fits, then book a call. It tells us what to prepare before we talk.
Three reasons to get in touch.
Our strategies are proprietary and run in-house, and we're open to conversations with investors. We go through the team, the engine, and where we are on a call.
Multi-venue market data on request, from the same capture our strategies run on.
Venues that want another professional maker on their books. Any quoting we do is for our own account.
A small team with real advantages.
Every venue captured against a single clock, so a spread between two venues is a real spread, not a timestamp artefact.
No third-party feed, strategy, or execution layer in the loop. We build each one into our proprietary engine, from connectors, normalization, capture, and research to strategies, order routing, and risk. So we know exactly what the strategy saw and did.
Captured streams replay in order, bit-for-bit. Strategies are backtested on the exact data they will trade on.
We build risk limits into the engine itself, so position, inventory, and loss limits apply before any order reaches a venue.
Native speed, memory-safe, no garbage-collection pauses. Latency stays predictable under load.
Perp DEXs and prediction markets alongside centralized exchanges, where spreads are wider and makers are fewer.
Where we point the engine.
We don't publish which venues we trade on.
Common questions.
What does Arbena do?
We build proprietary market-making and cross-venue arbitrage strategies, run for our own book. Every layer of the engine is in-house, including market data, research, strategy, execution, and risk.
Do you offer market making or arbitrage as a service?
No. Both are proprietary. We run them in-house for our own book, not for clients.
Are you open to investors?
Yes. We're open to conversations with investors. Book a call and we'll go through the team, the engine, and where we are.
Which venues do you trade on?
We target centralized exchanges, perp DEXs, and prediction markets. We don't publish the list.
Can I get your market data?
Yes, on request. We can share order book, trade, funding, and open interest data from the venues we capture, across centralized exchanges, perp DEXs, and prediction markets. Tell us what you need on a call and we'll say what we have.
What does market data cost?
It depends on the data types, venues, and history you need. We scope it with you on a call. There's no public price list.
Who is behind Arbena?
Arbena is a venture of CodeVotum, which architects data-intensive trading and IoT platforms. The engine is built on that engineering.
Let's talk.
Investor conversations, market data requests, and venue partnerships all start with a call.