TradeClanker

Traditional trading platforms revolve around opening and closing positions. TradeClanker revolves around the portfolio. Every trade becomes part of a permanent trading history, and holdings, cost basis, profit, and tax lots are all derived from that history rather than kept as running totals that can drift.

Most automated trading frameworks treat positions as the source of truth and bolt tax accounting on afterwards. Here the source of truth is an append-only ledger of orders and fills, so cost basis, lot selection, long-term-gains eligibility, and P&L are projections that can always be rebuilt and always agree.

The backtester replays saved market data through the same runtime that trades live, so a green backtest means something. Strategies are written in Lua or Python, and a strategy is free to hold, add to, or sell part of a holding rather than being forced into a sequence of entries and exits.

Highlights

• One engine, two data sources: live trading and backtesting share the same bot runtime and strategy code, and only the event source differs.
• An append-only event ledger. Orders and fills are immutable; basis, lots, gains class, and P&L are projections rather than mutable state.
• Strategies in Lua or Python, with the same functions and callbacks in either. Lua runs embedded and sandboxed, Python as an isolated subprocess.
• A web workbench for building a bot: edit the config as a form and the strategy as code, validate through the real engine, back-test the draft on screen, then run it on paper or live.
• Multi-jurisdiction tax lots, with FIFO, LIFO, and HIFO selection and a preference for long-term gains.
• Static Go binaries with no runtime dependencies. Postgres is optional for a durable ledger, and TimescaleDB optional on top of that.