FORGE mutates your strategy's code, backtests every variation, and keeps only what beats your baseline — validated on data no experiment ever touched before you see the result. No other bot does this.
Strategy Lab is built to iterate on trading logic specifically — with the validation discipline to back it up.
Every result is checked against a holdout slice of data no experiment ever trained on — the same discipline quants use to catch overfitting before it costs you.
Balanced, Conservative, Aggressive, or Consistency — tell Strategy Lab what kind of edge you're actually optimizing for.
Steer a run mid-flight with plain-English feedback. No need to stop and restart to redirect the AI.
Every real improvement is saved as a new strategy version the moment it's found — not just at the end of the run.
If the AI catches an actual bug in your strategy's logic, it fixes and adopts it immediately — verified to be a real code change, not a relabeled experiment.
Ask Strategy Lab to push a validated winning version straight to your live channel when the run finishes — the real validation pipeline runs either way.
From an existing strategy to a validated, improved version in four steps.
Choose any strategy you own and a scoring profile that matches your goal — return, low drawdown, or consistency.
Cap the number of experiments (up to 30) and see the exact wallet cost estimate before you commit to anything.
FORGE mutates the code, backtests each version, and keeps or reverts every change live as you watch.
The best version is tested on data it never touched, then saved as a new version — or auto-promoted live if you opt in.
Whether you're fine-tuning by hand today or automating the whole loop.
Already have a profitable strategy? Let AI search the parameter space for improvements you'd never find by hand.
Stop guessing at stop-loss percentages and indicator periods one backtest at a time.
Every capability here — start a run, inject guidance, validate, promote — is also exposed as an MCP tool for AI agents.
These numbers come straight from completed Strategy Lab runs — updated automatically, never hand-picked.
Wallet-funded, per-experiment pricing. Stop early and get a prorated refund on the rest.
See the full breakdown, including how prorated refunds work, on the pricing documentation.
Every Strategy Lab capability on this page — start a run, inject live guidance, cancel, validate, and promote a result — is exposed through TradeStaq's MCP server. Any MCP-compatible AI client can run an optimization on your behalf.
Pick a strategy you already own and let Strategy Lab find its best version.
Wallet-Funded • Pay Per Experiment • Stop Anytime