01 · Human
Design.
Frame the hypothesis. Define the constraints. Specify what counts as an edge and what would falsify it. The judgement is yours.
The Founding 20 is open
The agentic quant research platform.
You design the hypothesis. Agents run the research. You evaluate the results.
No signup needed to try. Pro is 30 days free, then $99/mo.
Full history. Full market. Every asset class.
US Stocks
Equities and ETFs.
US Futures
Index, energy, metals, agriculture.
Crypto
Major venues, 24/7.
Forex
Major and minor pairs.
Coming soon
Options
Equity and index options.
Coming soon
Prediction Markets
Event-driven contracts.
You design. Agents search. You evaluate.
You bring the hypothesis. Agents do the iteration. You decide what holds up.
02 · Cloud agents
Search.
Agents do the research toward goals you set: design experiments, refine the hypothesis, pursue what works. Walk-forward, out-of-sample, optimization, and meta-labeling are wired in. Every experiment is logged and reproducible.
03 · Human
Evaluate.
Read the results, decide what held up, and choose what to refine, retire, or deploy. AlphaMind keeps full context across runs so the comparison is real.
Cloud agents
Autonomous iteration, around the clock.
Define the hypothesis and the boundaries. Agents run the research, surface what works, and stop when the criteria are met.
Hypothesis-driven, not prompt-driven.
Agents work from a research brief: hypothesis, constraints, success criteria. They don't free-associate. The output is a research record: validated experiments and the evidence behind them, not a chat transcript.
Reproducible by default.
Every run is logged with seed, configuration, and validation results. Re-run any experiment, compare it to a sibling, or rebuild it later from the artefact alone.
Research only.
Agents don't trade. They don't deploy. They research and report. You decide what graduates from research to production.
Persistent context.
AlphaMind remembers your strategies, prior runs, and what you've validated. The next iteration starts from where the last one left off, not from a blank prompt.
Methodology
Rigorous by default.
Walk-forward, out-of-sample, optimization, and meta-labeling are wired into every backtest. Things that take serious effort to do correctly on your own are handled correctly by default.
Walk-forward, not single-window.
Strategies are re-fit on rolling windows and evaluated on the windows that follow. A backtest that holds up on a single split tells you what worked once. Walk-forward tells you whether it keeps working.
Out-of-sample, enforced.
Out-of-sample data is partitioned at the platform level, not at the strategy level. You can't accidentally peek. The validation surface is shared across runs so comparisons are fair.
Meta-labeling for ML.
Train models on user-defined labels without infrastructure overhead. Cross-validate, evaluate, and version models alongside the strategies that consume them.
Optimization, with discipline.
Parameter searches are scoped to the windows you specify and evaluated on the windows you don't. Overfitting is treated as a measurement problem, not a vibe.
Choose your plan.
Start free. Upgrade when you're ready. Early adopters lock in their rate for life.
Watch the 1-minute previewAlphaLab for firms
Set up for your whole desk.
A shared workspace for your quants, with consolidated billing and priority compute. Every firm gets hands-on onboarding, training for the team, and direct access to the people building the platform. We scope the rollout and the pricing to how your desk works.
Frequently Asked Questions
How is AlphaLab different from what I'm using today?
AlphaLab is the agentic quant research platform. You get a visual builder for speed, AlphaMind for scaffolding, and cloud agents that iterate on your strategies autonomously. Walk-forward, out-of-sample, optimization, and meta-labeling are wired in. Every run is logged and reproducible. The platform doesn't replace your judgment, it removes the plumbing.
I'm already a quant. What's here for me?
The loop. You design the hypothesis, cloud agents iterate on it, and you evaluate the results. Iteration runs around the clock while you work on the next idea. Walk-forward, out-of-sample, optimization, and meta-labeling are wired in by default, so the methodology stays correct without you having to police it.
Why wouldn't I just use Python?
You can. The question is what it costs you. Every strategy idea passes through a tax of setup, implementation, validation, and iteration before you know if it's any good. AlphaLab removes the tax. Cloud agents remove the waiting. If you'd rather build infrastructure than test ideas, stay in Python. If you'd rather spend your time on hypotheses, that's what AlphaLab is for.
Is AlphaLab a trading bot or a fully automated trading system?
No. AlphaLab is a research environment. Agents iterate on hypotheses, run experiments, and report. They don't trade. They don't deploy. You decide what graduates from research to production. A real edge is what you have that others don't, and anything sold to a crowd stops being an edge the moment the crowd buys it.
What markets and instruments are supported?
US stocks, US futures, forex, and crypto. Full markets across all four. Options and prediction markets coming soon.
How much does AlphaLab cost?
There's a free tier with core features and limited data. Paid plans start at $99/month for Pro and $299/month for Advanced during the early adopter program. Early adopters lock in their rate for life as long as they stay subscribed.