Inspiration
After losing 90% of our principal trading with AI-generated scripts, we found a gap between vibe coding and an executable trading environment.
What it does
We’re building Arcane, an AI Crypto Trading Bot Platform that lets anyone trade like a pro. Users describe trading strategies in natural language, and Arcane instantly generates a strategy document, deployable code, backtests, and one-click bot deployment.
How we built it
AI models: Gemini, Qwen, deepseek Agent tool call: Our AI leverages tools such as user demand collection, strategy design, coding, backtesting. RAG: Backtest scripts and deployment codes leverage Retrieval-Augmented Generation (RAG) technology. Generative UI: Parameter fields in the UI are dynamically generated by AI based on the specific requirements of each strategy. Context engineering: Robust prompts are essential to ensure the correct workflow, precise reasoning, and accurate memory retention.
Challenges we ran into
The live trades deployment parts is quite complex. It will takes time to build.
Accomplishments that we're proud of
We launched the MVP and gained users and positive feedbacks in 3 days.
What we learned
Ship as fast as you can.
What's next for Arcane
- launch live trades soon
- add more trading pairs, support perps
- more trading strategies such as DCA, GRID, etc.
- second-level backtest timeframe
Log in or sign up for Devpost to join the conversation.