About wickra-embed

wickra-embed runs the Wickra indicator math where there is no operating system and no heap: microcontrollers, FPGA soft-cores, HFT co-processors. Every update is O(1) with a bounded worst-case latency, uses fixed-capacity buffers (no allocation), and produces the byte-for-byte identical value the std wickra-core produces on a server — verified by a parity test suite.

What makes it different

• No OS, no heap. A #![no_std], allocation-free crate. Every indicator is a fixed-size state machine that lives on the stack or in a static — it runs on a microcontroller with a few KB of RAM.
• O(1), bounded latency. Every update is constant-time with a bounded worst case, the property a hard-real-time or HFT hot loop needs.
• Byte-for-byte identical. Each indicator matches its std wickra-core counterpart exactly, pinned by a parity corpus in CI, so the bare-metal path and the cloud path never disagree.
• Rust core + C ABI. A Rust #![no_std] crate (embed-core) plus a C ABI header and static library, so it drops into a Rust firmware project or any C / C++ toolchain.

What it ships

The v0.1 line ships five verified no-alloc indicators — Sma, Ema, Rsi, Atr and Roc — the distillation of the wickra-core math down to what runs with no heap. More follow as they are made allocation-free and parity-verified.

Why it exists

The full Wickra library assumes std and an allocator. wickra-embed is the bare-metal distillation of that same math, so a strategy can compute the exact same signals on a microcontroller at the edge as it does in the cloud — no reimplementation, no drift.

Open source

Released under the MIT OR Apache-2.0 license — permissive, OSI-approved, free for any use including commercial. Source, issues and releases on GitHub.

Disclaimer

wickra-embed is a software library, not a trading system, and is provided as-is with no warranty. It computes indicator values on-device; it does not give financial advice. Use it at your own risk.