Tag, version, promote, and monitor every strategy from draft to live deployment. Six lifecycle states, git-style lineage, real-time IC and drawdown alerts. One workbench, no spreadsheet.
Every strategy in your universe, counted, sorted, and observable. The numbers a portfolio manager would ask for, on the home screen.
From last week's drafts to year-old live veterans. One workbench, every state.
Draft, backtesting, paper, live, paused, archived. A clean promotion path, with no "it lives in a notebook".
Composed into the alpha factory, monitored for IC and drawdown in real time.
Roll back to any tagged commit, branch a variant, compare equity curves.
Every strategy carries the same metadata, the same lineage, the same state machine. Boring, in the best possible way.
Every parameter change is a commit. Branch a variant, compare equity curves, roll back to a tagged version. No more strat_v2_FINAL_final2.py.
Every strategy lives in exactly one state. Gates enforce promotion criteria, so you don't accidentally take a half-tested draft live because the row caught your eye in the wrong tab.
Every live strategy publishes its metrics to the dashboard and to your webhook. Set drawdown thresholds; auto-pause anything that breaches; ping Slack when something interesting happens.
Tag by family, instrument, regime, model, whatever your taxonomy demands. Pin folders for research vs production. Full-text search over names, descriptions, and parameter values.
Most quants manage their universe in a directory tree, a spreadsheet, and willpower. That works for ten strategies. It does not work for a thousand.
strat_v2_FINAL_final2.pyManaging scale is not a tooling luxury. It is the entire job. A signal factory without a workbench is just a graveyard. Strategy Management is the workbench.
No state machine to memorise, no Confluence page to consult. The promotion path is the same for every strategy, and it is visible right in the row.
Tag by family, instrument, regime, or any taxonomy you invent. Folders for research vs production. Full-text search across descriptions, parameters, and lineage notes.
Draft → backtesting → paper → live, with explicit gates. Set a paper-trade minimum (say, 30 days) before live is even an option. Every transition is reversible and logged.
Live IC, Sharpe, drawdown, and hit-rate, streamed off the engine and alerted via webhook. The strategy that earned a slot keeps it only as long as the numbers say so.
Every strategy in your universe, in one view. Filter by state, tag, family, owner. Click any row to open the full detail pane: versions, equity curve, parameter history, audit log.
Two threads to pull on next: read the playbook library these strategies are forked from, or learn the indicator math underneath.
Every strategy in your dashboard starts as a fork of one of these. Read the playbook once, then watch the variations multiply across your universe.
Browse the playbook →The primitives every managed strategy is built from. Formula, visual, edge cases, reference code. The math under your dashboard rows.
Read the indicators reference →The free tier ships with the C++23 engine, the full Strategy Management workbench, and YFinance + Dukascopy data. Six-state lifecycle, git-style versioning, and alerts, all from one strategy.
StratCraft is one of three engines we build. Same philosophy: open-source where possible, performance-first, no lock-in.
C++23 FIX protocol engine. 3× faster than QuickFIX with zero-allocation design. Production-grade message handling for venues that don't forgive a missed heartbeat.
Learn more →The networking and identity layer for AI agents: discovery, naming, encrypted relay, and autonomous cooperation. The wires beneath the agent economy.
Learn more →