|
1 | | -# QuanuX: The Great Equalizer |
2 | | -*Biological Sovereignty in Algorithmic Execution* |
3 | | - |
4 | | -**QuanuX** intends to become a complete, end-to-end ecosystem for high-frequency execution, AI-driven strategy generation, and institutional-grade observation. Built for those who refuse to be limited by conventional infrastructure, QuanuX bridges the gap between AI-driven research and ultra-low latency execution. From bare-metal C++ cores to reactive cockpits, QuanuX provides the tools required to compete natively at the edge of the market. It is the absolute Equalizer, demanding nothing but intellect and effort. |
| 1 | +# QuanuX: Personal Quantitative Trading Lab |
| 2 | +*AI Ready Self Hosted Multi-Node Research, Development, Deployment and Portfolio Managemnt Framework & Platform* |
5 | 3 |
|
6 | 4 | [](https://www.gnu.org/licenses/agpl-3.0) |
7 | 5 | [](https://isocpp.org/) [](https://www.python.org/) [](https://www.rust-lang.org/) [](https://cython.org/) [](https://go.dev/) |
8 | 6 | [](https://www.typescriptlang.org/) [](https://tailwindcss.com/) [](https://react.dev/) [](https://vitejs.dev/) [](https://www.anaconda.com/) |
9 | 7 |
|
| 8 | +**QuanuX** a Free Open Source Software initiative (FOSS) intends to become a complete, end-to-end ecosystem for high-frequency execution, AI-driven strategy generation, and institutional-grade observation. Currently, QuanuX has integrations for several brokers and prop firms utilizing Python and C++ APIs, including FIX adapters for qualified clients. Built for those who refuse to be limited by conventional infrastructure, QuanuX bridges the gap between AI-driven research and ultra-low latency execution. From bare-metal C++ cores to reactive cockpits, QuanuX is developing the tools required to compete natively at the edge of the market. It is the absolute Equalizer, demanding nothing but intellect and effort. |
| 9 | + |
| 10 | +Each module is accompanied with detailed **SKILL** files. Connect any AI hosted locally or remote over API via our MCP server -- agents are trained throughout the codebase on the purpose for each module, the workflow design, coding patterns, and how to integrate with other modules and/or extend the codebase. For example, our integration with Figma MCP allows one to design a feature visually using only prompts with Figma Make AI generating frontend React code. Then, import the React design into QuanuX client directories via AI prompts where hooks and mock data calls are rewritten into QuanuX compliant methods and pointed to the actual backend data source in their cluster. |
| 11 | + |
| 12 | +Nodes are intended to be hosted on Linux servers with cross platform clients and our CLI terminal client 'quanuxctl' allowing users to connect to owned or leased remote servers, where they can then be conditioned into QuanuX nodes -- OS and security customization with RSA (or other) handshake and NATS private network invitation. Once a node is conditioned, it can then be repurposed into a specific node genre (i.e. regular execution, FIX execution, NIC/Solarflare, observation, statistics, etc.). |
| 13 | + |
| 14 | +This is the foundation of the "Russian Doll" concept. Rather than call this a "block" system where pieces are stacked upon one another, in QuanuX each layer is stacked *inside* one another. The outermost layer is the basic server/networking controller, then via quanuxctl the preferred node type can be deployed within that node i.e. an execution node or statistical engine, etc. This is endlessly customizable, layer by layer, with the core of the Russian Doll 'the baby' always carrying the core purpose of the node, be it a trading strategy, HFT engine, statistical computation engine, audit & reconcilliation, risk observation, backtesting, replay, etc. The baby is protected at the core, all surface receptors perfectly insulated (hence heavy biological reference in documentation). |
| 15 | + |
| 16 | +A great deal of effort has gone into 'bare metal' execution nodes with underlying control of processor cores/threads and increasing effort is being applied in execution nodes intended for NIC/Solarflare Direct Exchange/Broker Access (DMA) which entirely bypass the Linux OS. The purpose for this is typical SCRUM rationale -- do the hardest part first and everything else falls into place from there. That said, most will never need this functionality and much of the more basic order/data functionality (i.e. REST API Python & C++ non-FIX -- those not requiring conformance testing and independent confirmation to trade) is ready to go. |
| 17 | + |
| 18 | +The AI Strategy Lab is a core feature in development. The dream **Python --> Cython --> C++** is becoming a reality. This will allow traders to work in Python which opens up the world of tools most of us use on a daily basis in a language that is easy to be comfortable with and fast for idea iteration. The Anaconda integration allows environments to be tightly managed and serves as a "source of truth" for package management and QuanuX binaries while enabling seamless transition between QuanuX and Machine Learning/AI tools like PyTorch. Launch QuanuX right from Anaconda Navigator when you install the channel! (More details soon...) |
| 19 | + |
| 20 | +QuanuX backtesting software is coming along quickly and is being built around Databento L3 historical datafeed which QuanuX hopes to offer directly very soon at a massively affordable rate versus what you can find anywhere else. QuanuX is also exploring "one click deploy" with DigitalOcean so traders who do not possess their own server cluster infrastructure can let QuanuX and DO take care of it for them -- this is a lot more affordable than you may expect -- and will never feature a markup over DO's stated rates. Although not the same as having a bare metal server connected by a 3' Cat6 cable at the exchange datacenter, being closer to the exchange with 100% uptime and much lower latency should result in edge for prop and retail traders who aren't building strategies based on sub millisecond signals. |
| 21 | + |
| 22 | +This is a large, ambitious project coming very close to its first release. Your support will absolutely help bring QuanuX v0.0.1 LIVE more quickly -- financial via GitHub Sponsors or code contributions, everything helps! |
| 23 | + |
10 | 24 | --- |
11 | 25 |
|
12 | 26 | ## ⚡ Pillar I: The Ultra-Low Latency Core (59ns Execution) |
|
0 commit comments