Event-sourced trade execution system that turns trading decisions into venue orders, durable execution state, and auditable outcomes.
Praxis unifies order routing, venue communication, lifecycle management, and recovery around a single append-only Event Spine. Persist-before-send order submission means no trade is ever lost to a crash, and startup replay rebuilds state from the durable event log. The current repository is a partial implementation of RFC-4001 (Praxis issue #1); where the RFC and the code diverge, the code is authoritative.
Praxis is not:
- the strategy or decision layer
- a generic multi-venue execution platform
- the full RFC-4001 system as written
- a live-trading-hardened production deployment
In the wider Vaquum architecture, Origo sits upstream as the data layer and Limen as the research engine. Nexus turns research outputs into decisions, Praxis executes and tracks those decisions on the venue, and Veritas sits downstream for oversight.
- Append-only SQLite Event Spine with a tamper-evident SHA-256 hash chain
- Persist-before-send order submission with deterministic client order IDs
- Event-sourced paper trading on Binance Spot testnet
- Support for market, limit, and IOC orders, with stop-loss and take-profit expressed as OCO legs
- Per-account execution routing with isolated credentials and independent account runtimes
- Durable tracking of fills, open orders, closed orders, and per-trade positions
- Per-account double-entry ledger with balances and per-trade realized profit and loss
- Terminal trade outcomes for fills, aborts, cancels, and deadline expiry, delivered through async callbacks for Manager integration
- Venue filter loading and pre-submission validation against Binance trading rules
- Startup replay, venue reconciliation, and WebSocket recovery for crash-safe state rebuilds
- Walk-the-book execution and arrival slippage estimation, surfaced per fill in structured logs
- Binsim in-process Binance simulator and deterministic replay harness for offline runs
Two execution gaps are tracked for Q3/26:
- Standalone stop, stop-limit, and take-profit submission through the venue adapter (Praxis issue #171)
- Execution and arrival slippage fields on
TradeOutcome(Praxis issue #172)
The first runnable path is the local test suite followed by an event-sourced Trading boot against a fresh Event Spine.
- Clone the repo and install the package with dev dependencies:
git clone https://github.com/Vaquum/Praxis.git
cd Praxis
uv pip install -e ".[dev]"Praxis requires Python >=3.12 and installs its venue, persistence, and observability dependencies from pyproject.toml, including vaquum-nexus pinned from GitHub; the package is not published on PyPI. The dev extra adds the test toolchain, and pip install -e ".[dev]" with python -m pytest works equally if uv is not available. No venue credentials are needed for the install, the test suite, or the first boot; report security concerns through the Vulnerabilities section below.
- Run the default test suite:
uv run pytest- Boot the event-sourced runtime against a fresh Event Spine, which writes
event_spine.sqliteto the current directory:
uv run python - <<'EOF'
import asyncio
import aiosqlite
from praxis import Trading, TradingConfig
from praxis.infrastructure.event_spine import EventSpine
async def main() -> None:
async with aiosqlite.connect('event_spine.sqlite') as conn:
spine = EventSpine(conn)
trading = Trading(config=TradingConfig(epoch_id=1), event_spine=spine)
await trading.start()
await trading.stop()
asyncio.run(main())
EOF- Optionally verify Binance Spot testnet access with credentials in
.envor shell exports:
echo 'BINANCE_TESTNET_API_KEY=your_key' >> .env
echo 'BINANCE_TESTNET_API_SECRET=your_secret' >> .env
uv run pytest tests/testnet/ -v -o 'addopts='That path verifies the implementation as it exists today: local execution logic by default, then Binance Spot testnet connectivity when credentials and network access are available. Beyond the first run, praxis.launcher boots the full per-account runtime with Nexus decision wiring, and the praxis.binsim package serves an in-process Binance-shaped venue for offline paper trading.
Praxis is research software. Paper-trading, simulation, and replay outputs are not investment advice, trading advice, live-execution guarantees, regulatory approval, or a promise of future performance. Past performance is not predictive, and trading digital assets can result in total loss of capital.
- Start with the documentation hub in docs/README.md
- See the target design in RFC-4001 and known constraints in TechnicalDebt.md
- Start with Setup-And-Verification.md and the runtime walkthroughs in Trading.md and Trade-Lifecycle.md
- Use Event-Spine.md and Trading-State.md for persistence and projections
- Use Execution-Manager.md for the execution core and Venue-Adapter.md for venue integration
- Strengthen recovery understanding with Recovery-And-Reconciliation.md and Health.md
- Analyze execution quality with Slippage-And-Order-Book.md and Trade-Outcomes.md
- Simulate the venue offline with Binsim.md and verify connectivity with Binance-Spot-Testnet.md
- Read the runtime entry points in praxis/trading.py, praxis/launcher.py, and praxis/trading_inbound.py
- Contribute through the Developer docs
Contribution starts from the gap between RFC-4001 and the current implementation, through open issues and docs changes. Before contributing, read the code first — the current implementation is the source of truth where it conflicts with the RFC — and start with the Developer docs.
Use GitHub issues for support requests and scope questions.
Report vulnerabilities privately through GitHub Security Advisories. Do not report vulnerabilities through public issues.
Published work should cite:
Vaquum Praxis [Computer software]. (2026). Retrieved from GitHub.