Custom Studies · Native C++
ProThe terminal ships with its own C++ compiler. Open the built-in Study Editor, write your study against a single C-ABI header, and press Compile & Load — it builds and drops straight onto the live chart as a first-class study beside the built-ins. Full native speed, direct access to every bar and order-flow series, and crash isolation so a faulting study never takes the terminal down with it.
Bolt-on scripting languages were always a compromise. Real desks extend the platform in the platform's own language. Sharpnel hands you that same lever — native C++, an ABI you can build against for years — bundled with Pro, not sold as a five-figure add-on.
What you get
Your study compiles against a single pure-C header — no Qt, no Skia, no STL crossing the boundary, just POD structs and function pointers. The ABI is versioned with additive-only evolution, so the study you build today keeps loading as the terminal updates.
Studies run in-process at native speed, computed off the render thread — a heavy study can't stutter the chart. They read the same series the terminal draws — open / high / low / close / volume, buy and sell volume, delta, price source. No round-trip, no transpile, no interpreted-runtime ceiling.
The host owns every output buffer and plants a canary around it. A study that faults or writes out of bounds is fault-disabled for the session — the terminal stays up by design.
Up to 16 subgraphs per study with line, histogram, area, dots, band, and step styles, plus reference levels and drawings. Your study looks exactly like SMA, EMA, RSI, or Delta — because those ship on the same SDK.
Declare int, double, bool, enum, color, and price-source inputs; the terminal renders the settings dialog for you. Studies recompute only when the data or your inputs actually change — not on every repaint.
Studies compute on whatever feed you connect — bring your own broker for direct L1 / L2 on futures, equities, or crypto. The study you write runs on the data you own, not on a relayed copy.
What you'd do with it
You've carried a study for a decade. Rebuild it in C++ against the SDK, drop the DLL in, and it's a native study on the chart — the same speed as the built-ins, none of the rewrite-it-in-someone-else's-DSL friction.
Per-bar delta divergence, absorption scoring, footprint-derived signals — compute-heavy work a scripting layer chokes on runs at full speed here, because it is native code, not an interpreted layer on top.
Compile a suite of in-house studies into DLLs your whole desk loads. The ABI is stable, so last quarter's build keeps working after the terminal updates — your library is an asset, not a maintenance treadmill.
Included with Pro ($125/mo) — that's about 1/16th the cost of an institutional desk terminal.
Common questions
Nothing — the C++ compiler ships inside the terminal. Open the built-in Study Editor, write your study against the single SDK header, and press Compile & Load (Ctrl+Enter): it builds and drops straight onto the active chart, zero setup on a fresh install. Prefer your own toolchain? Studies also build externally — MSVC, x64 Release, dynamic CRT — so a desk can compile in its own CI and drop the DLL in the studies folder, where Add Study lists it next to the built-ins.
No. The ABI is frozen and versioned. New capabilities are added additively behind reserved fields, so a v1 study keeps loading. A genuinely breaking change bumps the major version and the host refuses the mismatch cleanly rather than crashing.
Native C++ studies run in the Sharpnel desktop terminal, bundled with Pro. The free Explorer tier lets you evaluate everything before subscribing.
No. Studies you compile are yours — keep them private, share the DLL with your desk, distribute them however you like. No marketplace tax, no lock-in, no phoning home.
