ReFi.Trading offers self-custodied trading agents that execute trades directly from user brokerage accounts, ensuring full asset control. The platform uses reinforcement learning models and zero-knowledge proofs for automated, disciplined trading with verifiable risk parameter adherence.
Funding
Funding not disclosed
Founders
Product
Problem
Manual trading is susceptible to emotional biases, leading to suboptimal execution and discipline. Deploying sophisticated reinforcement learning (RL) strategies for trading is complex and requires significant technical expertise, creating a barrier for many retail and quant-curious individuals.
Solution
ReFi.Trading provides self-custodied trading agents that execute trades directly from a user's existing brokerage account, ensuring assets remain under full control. The platform enforces risk parameters using zero-knowledge proofs, verifying compliance without revealing sensitive strategy details. This enables automated, disciplined trading by leveraging advanced RL models and a secure, transparent infrastructure. Users can connect their brokers, define risk limits, and launch autonomous trading agents with minimal setup.
Target Audience
The platform targets retail traders seeking to overcome emotional biases and quant-curious individuals looking for a simplified way to deploy and manage RL-based trading strategies.
Features
- Self-custodied trading agents that execute directly from linked brokerage accounts.
- Zero-knowledge (zk) proofs for pre-trade risk verification, ensuring adherence to defined parameters.
- Reinforcement learning (RL) models for strategy development and execution.
- Agent management console for launching, stopping, and monitoring trading activities.
- Secure connection to brokerage accounts, maintaining user asset control.
- Transparent audit trails for all executed trades and risk verifications.
- Support for connecting to multiple brokerage platforms.
- API for programmatic strategy development and integration.