No-Code Platform for Backtesting Crypto Trading Strategies
Hey indiehackers,
I've built a platform https://cryptoquant.dev that lets traders easily backtest crypto trading strategies without any coding required.
How It Works
After exploring code-based solutions like https://github.com/jozefRudy/trading-rule-parser, I implemented https://developers.google.com/blockly for a visual programming approach. This gives users flexibility while preventing logical errors by only allowing blocks to connect in ways that make sense.
The platform offers both backtest and live trading modes. I've written a guide on building a basic strategy https://cryptoquant.dev/resources/blog/improve-buy-and-hold-btc that outperforms buy-and-hold for BTC.
Current Features
- Visual strategy builder (no coding required)
- Backtesting against historical data
- Live trade notifications
- Portfolio approach with fixed universe selection (e.g., top 10 cryptocurrencies)
- Coverage of 400+ cryptocurrencies on Binance
- Minute-by-minute data going back to February 2018
Future Roadmap
I'm considering several expansion paths:
1. News sentiment analysis - Classifying financial news by market impact/direction to create rules for exiting positions after significantly negative news
2. Dynamic universe rotation - Allowing universe selection based on previous performance metrics rather than fixed assets
3. Social media monitoring - Alerts based on social network activity for specific coins
About Me
I'm a former portfolio manager turned solo builder, deliberately choosing the no-financing journey. I've built feature-request feature https://cryptoquant.dev/resource/feature-requests directly into the platform.
Would love to hear your thoughts, feature suggestions, or any questions about the platform!
Great work, I am just getting started on no code Ai automation
there are some solutions in place that actually combine no-code (as visual programming) and ai in trading, e.g. composer dot trade (cannot post links still). Thought about that as well, but that's another rabbit hole, since how to guarantee that ai NEVER hallucinates and outputs valid code-blocks connected together is another story. So don't envy you in that respect :)
Great work, jozef! The visual Blockly builder looks super intuitive for non-coders. Curious ,how are you managing minute-by-minute data latency, and any plans to add other exchanges beyond Binance?
actually i found there are exchanges that have many more assets, e.g. coinex or mexc (around additional 2,000 cryptocurrencies, opposed to binance of 400). Maybe it would make sense to add spot assets (all just against USDT) from top 10 exchanges or so, would it then be a more useful product?
Thank you.
What do you mean by how I manage minute-by-minute latency?
There are multiple reason why live strategies will perform well even on minute-on-minute data, most important are columnar storage, and performant dotnet core framework (with f#) with possibility of horizontal scaling (architecture) empowered by redis streams.
I wrote a blogpost about the architecture I use (please find it in my blog on website since I am not allowed to post links 🤦🏻 here yet ). I consume minute-by-minute data in real time directly from binance which has around 400 cryptocurrencies.
I believe binance has probably most of cryptocurrencies of what others offer, but if you think there are some instruments that are missing and you have some suggestions and you think they are important, I am happy to get some tips. It is always a question of priority, of adding new features, hence adding new exchanges is not planned for now, but if I can see a significant reason, it can surely be considered. Do not hesitate to post a suggestion here (on cryptoquant it is possible directly in resources -> request feature, but again, I am not allowed to post links here 🤦🏻 yet).