wickra-terminal 0.1.4

---

One core. Ten languages. Two renderers. A streaming trading terminal built on the Wickra core — live charts, order-book, tape and 497 streaming indicators — with a native TUI and a Web front-end as a second renderer of the same logic, driven by the same config.

# The native TUI over a live Binance feed. No config file, no account, no key.
cargo run -p wickra-terminal -- --source live:binance:BTC/USDT

# Or a deterministic synthetic feed, if the network is not the point right now.
cargo run -p wickra-terminal -- --source synth:1

Every one of the 514 indicators and bar builders Wickra ships is reachable: 497 through the indicator registry, 6 as profiles, 10 as alternative bar types, and the footprint through the panel that renders it. Three surfaces rather than one list, because a reading, a histogram and a bar are not the same shape of answer — see docs/INDICATORS.md.

▶ Live demo: the Wickra library's own 514 indicators over real Binance market data, computed live in your browser — live.wickra.org · zero backend, powered by wickra-wasm.

Part of the Wickra ecosystem: the same data-driven core and ten-language binding surface also power wickra-exchange, wickra-backtest, wickra-screener and 20 more — see the full list.

The heart is a single data-driven core, wickra-terminal-core: it folds market events into an O(1) AppState and turns panels into view-models (values, series, colours) — never renderer commands. The TUI maps a view-model to a ratatui widget; the Web app maps the same view-model to a canvas draw. One logic, N front-ends.

Data arrives through the DataSource trait, an activatable module:

• Live — the wickra-exchange connectivity layer over the ten largest venues. Native builds only: that client is native code with a socket and an HTTP stack, so the browser opens the venue's public stream itself and feeds a Manual source instead. That bridge is hand-written and reaches Binance spot — see docs/RENDERERS.md.
• Replay — a recorded feed with a time-machine seek: the whole event list is kept, so Seek rewinds and re-folds state deterministically. It reads no files and holds no engine, which is why it runs in the browser too.
• Synth — a deterministic synthetic feed for demos and tests.

The core is exposed as a JSON-over-C-ABI data API (Terminal::command_json) in Rust, Python, Node.js, WASM, C, C++, C#, Go, Java and R — so a developer in any language builds their own front-end on the same core.

Why this shape

Most terminals are an application with a data layer inside. This one is a data layer with two applications outside it, and that difference is what the repository is actually about.

• The renderer is not where the logic lives. Panels emit view-models — values, series, rows — never draw calls. The TUI maps them to ratatui and the browser to a canvas and some tables, and neither can diverge in behaviour because neither has any.
• The boundary is data, not an API. A config JSON in, a command JSON in, a frame JSON out. That is why 497 indicators became reachable from ten languages without a line of binding code, and why a third renderer needs no core change.
• One core, checked across ten languages. Not "ports that should agree" — one Rust core behind a C ABI, with every binding asserting the same recorded feed produces the same frame byte for byte.
• O(1) per event, bounded everywhere. State folds forward and never recomputes over history; the tape, the price series and each indicator's series are all capped. A feed that never stops does not grow the process.
• Exact where it matters. Prices and quantities are Decimal through the market layer, converted to f64 only at the view-model edge, where they are drawn rather than compared.

Status

0.1.4 — the current release. The core, both renderers, all ten language bindings, the indicator registry, the runtime source/symbol toggle, the panel set, the byte-exact golden corpus, property and fuzz tests, benchmarks and one runnable example per language are in place and green across the full CI matrix (10 languages x 3 OS). ROADMAP.md has what is done, what is open and what is not planned.

Read-only. The terminal renders market data. It places no orders, holds no credentials and keeps no position — the live source connects to public endpoints with empty credentials. Execution is not a flag that is off; it is a layer that is not built. See THREAT_MODEL.md.

Documentation

• Architecture — the core, the renderer split, the data-driven boundary.
• docs/INDICATORS.md · docs/PANELS.md · docs/SOURCES.md · docs/RENDERERS.md · docs/STREAMING.md · docs/Cookbook.md.
• ROADMAP.md · BENCHMARKS.md · SECURITY.md.

Quickstart

# Native TUI renderer over a live Binance feed:
cargo run -p wickra-terminal -- --source live:binance:BTC/USDT

# Or a deterministic synthetic feed (no network):
cargo run -p wickra-terminal -- --source synth:1

# Or replay a recorded feed. `replay:` takes the JSON itself, not a path, so
# a recorded file is passed through the shell:
cargo run -p wickra-terminal -- --source "replay:$(cat golden/replay/basic.json)"

# Anything beyond one source -- panels, indicators, timeframe -- comes from a
# config file, which overrides `--source`:
cargo run -p wickra-terminal -- --config my-terminal.toml

# Record the session, and reach further back on the first subscription. Both
# apply on top of `--config`: a stored layout is a layout, and these two are
# decisions about this run.
cargo run -p wickra-terminal -- --source live:binance:BTC/USDT --record 50000 --backfill 500

The three --source shorthands are synth:<seed>, live:<venue>:<BASE/QUOTE>[:<market>] and replay:<json> — the last one taking the recorded events inline rather than a filename. A live source opens the spot book unless a market follows the symbol: spot, usdm, coinm or margin, so live:binance:BTC/USDT:usdm watches the USD-margined perpetual. See docs/SOURCES.md for what each one does and docs/Cookbook.md for worked config files.

The recorder can also be started while the terminal is running — r in either renderer takes a capacity, and an empty answer stops it — so keeping a session does not mean having decided to before it began.

Renderers

Renderer	Where	How
TUI	native terminal	crates/ui-tui (ratatui), cargo run -p wickra-terminal
Web	browser	web/ (Vue) over bindings/wasm, cd web && npm run dev

Both consume the identical Frame of view-models from wickra-terminal-core, and both read the same keymap out of the same config. Where they differ they differ in idiom rather than in capability — a prompt in the terminal is a focused field in the browser — with one exception worth knowing before you choose: live market data reaches ten venues natively and one in the browser. docs/RENDERERS.md has the rest.

The Web row has a prerequisite the TUI row does not: web/ depends on file:../bindings/wasm/pkg, which wasm-pack produces and no clone contains, so npm install fails until it has been built once.

( cd bindings/wasm && wasm-pack build --target web )
cd web && npm install && npm run dev

Install

Binding	Install	Example
Rust (TUI binary)	cargo install wickra-terminal	examples/rust/src/main.rs
Python (PyO3)	pip install wickra-terminal	examples/python/synth_terminal.py
Node.js (napi-rs)	npm install wickra-terminal	examples/node/synth_terminal.js
Browser / WASM	npm install wickra-terminal-wasm	web/ — the Vue renderer
C / C++ (C ABI)	header + library, see bindings/c	examples/c/synth.c
C# (C ABI)	dotnet add package WickraTerminal	examples/csharp/Program.cs
Go (cgo, C ABI)	go get github.com/wickra-lib/wickra-terminal-go	examples/go/synth_terminal.go
Java (FFM, C ABI)	Maven Central org.wickra:wickra-terminal	examples/java/SynthTerminal.java
R (.Call, C ABI)	install.packages("wickraterminal", repos = "https://wickra-lib.r-universe.dev")	examples/r/synth_terminal.R

examples/README.md is the cross-language index; every one of them drives the same core through the same JSON commands.

C and C++ link the C ABI directly; the header is generated and committed at bindings/c/include/wickra_terminal.h.

Use in any language

The same Terminal handle — construct from a JSON config, drive with command(json) -> json, read version — is reachable from every binding:

import json
from wickra_terminal import Terminal

term = Terminal(json.dumps({"sources": [{"Synth": {"seed": 1}}]}))
term.command(json.dumps({"type": "Subscribe", "source": 0, "symbol": "BTC/USDT"}))
frame = json.loads(term.command(json.dumps({"type": "Tick"})))
print(frame["panels"][0])          # the chart panel's view-model

Benchmarks

The frame budget is dominated by the terminal's own CPU work — folding events and building view-models — not by rendering. Criterion medians on a Windows x86-64 laptop, single-threaded, with the default two-indicator overlay:

Path	Median	Throughput
Fold one trade into state	142 ns	~7.0 M/s
Apply an L2 depth diff	107 ns	~9.3 M/s
Build all five panels' view-models	8.9 µs	~113 K/s
One full tick (poll + fold + build)	9.7 µs	~103 K/s
The same tick across the FFI boundary	17.7 µs	~56 K/s

A full tick costs about ten microseconds, so the core sustains a hundred thousand of frames per second — far above any renderer's budget, which is the point of the O(1) fold. The indicator count is a direct multiplier: those numbers are two indicators, and the registry offers 497. See BENCHMARKS.md.

Project layout

crates/wickra-terminal-core   the data-driven core (DataSource, AppState, panels → view-models)
crates/ui-tui          the native TUI renderer (bin: wickra-terminal)
crates/wickra-terminal-bench  criterion benchmarks
bindings/{python,node,wasm,c,go,csharp,java,r}   the ten-language surface
web/                   the Vue/Vite Web renderer over the WASM binding
golden/                recorded feeds + byte-exact expected frames (cross-language parity)
fuzz/                  cargo-fuzz targets (feed_event, state_fold, view_model, config_parse)
examples/              one runnable example per language
docs/                  indicators, panels, sources, renderers, streaming, cookbook

Building everything from source

# Rust core + tests + lints
cargo build --workspace
cargo test  --workspace --all-features
cargo clippy --workspace --all-targets --all-features -- -D warnings
cargo run -p wickra-terminal -- --source synth:1

# Python binding (requires a Rust toolchain + maturin)
( cd bindings/python && maturin develop --release ) && pytest bindings/python/tests -q

# Node binding (requires @napi-rs/cli)
( cd bindings/node && npm install && npm run build && npm test )

# WASM binding (requires wasm-pack) — two targets: `web` for the renderer,
# `nodejs` because it is the only form `require()` loads
wasm-pack build bindings/wasm --target web
wasm-pack build bindings/wasm --target nodejs --out-dir pkg-node
node --test bindings/wasm/tests/*.test.cjs

# C ABI (cdylib + staticlib + the generated header)
cargo build -p wickra-terminal-c --release
.github/scripts/check-cbindgen.sh            # header still in sync?

# C# binding (requires the .NET 8 SDK; links the C ABI above)
dotnet test bindings/csharp/WickraTerminal.Tests

# Go binding (requires a C compiler for cgo; links the C ABI above)
( cd bindings/go && go test ./... )

# Java binding (requires JDK 22+ and Maven; links the C ABI above)
mvn -f bindings/java/pom.xml test

# R binding (requires a C toolchain / Rtools; links the C ABI above)
R CMD INSTALL bindings/r && Rscript bindings/r/tests/run_tests.R

# Web renderer (requires the `web` wasm target above)
( cd web && npm install && npm test && npm run build )

Testing

Every language asserts its own output against golden/ — one recorded feed and the frame it must produce, compared byte for byte. Rust, Python, Node, WASM, Go, C# , Java and R each run it in their own suite; C and C++ run it through the C ABI under ctest. Ten languages, one file: that is what makes "the same core everywhere" checkable rather than asserted.

Alongside it: property tests over the fold invariants, four cargo-fuzz targets across the parsing paths, a conformance suite pinning the trait shapes, a registry suite that constructs and drives all 497 indicators, and a test that extracts every example from this README and the docs and runs it.

cargo test --workspace --all-features
python scripts/check_binding_surface.py     # every binding matches the C ABI header

Requirements

• Rust ≥ 1.88 to build the wickra-terminal TUI (ratatui pulls instability/darling), and to build the Node binding. The library crate wickra-terminal-core keeps the workspace MSRV of ≥ 1.86.
• Renderer/binding toolchains as needed: Node ≥ 22, Python ≥ 3.9, a C toolchain, .NET 8, JDK 22+, Go 1.23, R ≥ 4.1 — see each bindings/<lang>/README.md.

Ecosystem

Part of the Wickra family — each one a data-driven core with a CLI and the same ten-language binding surface:

• wickra — main library (Rust core + Python / Node.js / WASM bindings + a C ABI for C / C++ / C# / Go / Java / R)
• wickra-playground — a polyglot strategy playground: one StrategySpec live side by side in Python, Rust, JS and Go, entirely in the browser
• wickra-exchange — unified market-data + execution across ten crypto exchanges
• wickra-backtest — event-driven backtester over the Wickra core
• wickra-screener — parallel multi-symbol screening over 514 streaming indicators
• wickra-xray — market-microstructure explorer: footprint, order-book heatmap, liquidation map, funding/OI divergence
• wickra-radar — perp-universe alert radar: OI delta, funding flip, book imbalance, liquidation clusters, OI/price divergence
• wickra-copilot — local market copilot grounded in real order-book, liquidation and funding microstructure
• wickra-shazam — match an asset's current microstructure fingerprint against its entire history
• wickra-benchmark — reproducible, golden-verified benchmark suite — recompute any (strategy, dataset, report) in ten languages and confirm it byte-for-byte
• wickra-strategy-ci — Jest for trading strategies: golden-pin the report, catch regressions in CI, property-test against fuzzed data
• wickra-verify — confirm or refute a claimed backtest report against its strategy and data, in ten languages
• wickra-proof — Proof-of-Backtest: deterministic (spec, data) → report + blake3 hash, recomputable byte-for-byte in ten languages
• wickra-zk — prove a backtest zero-knowledge — on-chain-verifiable performance without revealing the data or the strategy
• wickra-impact — the backtester that knows you would have moved the market: agent-based fills on the real historical L2 order book
• wickra-darwin — evolutionary strategy search at millions of backtests per second, mutating and crossing JSON specs across the 514-indicator space
• wickra-gym — a Gymnasium-compatible, microstructure-aware backtest environment with O(1) steps for deterministic RL rollouts
• wickra-feature-store — OHLCV and microstructure streams into ML-ready feature matrices over 514 O(1) streaming indicators
• wickra-genome — a vector database of the whole market: every asset a 514-dim live vector, for similarity search, clustering and anomaly detection
• wickra-timemachine — scrub the whole market like a video — every symbol, full order book, rewound to any moment via deterministic re-fold
• wickra-synth — deterministic synthetic market microstructure: OHLCV, order book, trades and funding from a single seed
• wickra-compile — compile a strategy spec into a standalone deployable: a WASM module, a self-contained binary, or a no_std artifact
• wickra-embed — allocation-free, no_std streaming indicators for bare-metal and HFT, byte-for-byte identical to the core
• wickra-pico — the O(1) indicator core running bare-metal on a $5 Raspberry Pi Pico — the LED blinks on the EMA cross

Docs at docs.wickra.org; the marketing site and in-browser demo at wickra.org.

Contributing

See CONTRIBUTING.md and CODE_OF_CONDUCT.md. Commits are signed and in English; open a PR against main.

Security

See SECURITY.md and THREAT_MODEL.md. Report vulnerabilities privately — never in a public issue.

License

Dual-licensed under either MIT or Apache-2.0, at your option.

Disclaimer

This software is provided "as is", without warranty of any kind. It is a research and engineering tool, not financial advice. Trading carries risk of loss. Run in paper mode and against exchange testnets, and review the code before risking real capital.

---

Built on Wickra. If it saved you time, the cheapest way to say thanks is to ⭐ the repo.