Pricing
Pay for compute. Cancel anytime.
Three plans. No seat licensing, no usage cliffs, no vendor lock-in.
Launch pricing — these allocations are locked through 2026-12. We’re building the agent-native default for backtesting and want you to iterate freely without budget anxiety. Subject to revision in 2027.
Free $0 | Trader $19/mo | Quant $39/mo | |
|---|---|---|---|
| Monthly credits | 500 (rolling 30-day refresh) | 30,000 | 300,000 |
| What that means in practice | ~10 short backtests for evaluation | ~550 single 1y minute backtests, or 30+ medium sweeps | Walk-forwards + tick mode + heavy sweeps — designed to not feel limiting |
| Crypto symbols | BTC, ETH, SOL | Top 50 USDT pairs | Top 50 USDT pairs |
| FX symbols | — | 16 majors (tick history from 2003+) | Same — tick mode unlocked |
| US stocks | — | Top 200 + 1m intraday + EOD | Same |
| ETFs | — | ~50 curated (sector, bond, commodity, vol, intl, factor) | Same |
| Fundamentals | — | — | Earnings, ratios, statements (DOW 30 today) |
| Minute mode | ✓ | ✓ | ✓ |
| Hybrid mode (adverse-fill realism) | — | ✓ | ✓ |
| Tick mode (full event-driven) | — | — | ✓ (BTC, ETH, SOL) |
| Parameter optimization | — | up to 50 combos / sweep | up to 1,000 combos / sweep |
| Walk-forward analysis | — | — | ✓ |
| Persistent strategy library | — | ✓ | ✓ |
| Public result sharing | — | — | ✓ |
| Strategy comparison tool | — | — | ✓ |
| Funding-rate aware perp backtests | — | — | ✓ |
| API keys | 1 | 3 | 10 |
| Concurrent backtests | 1 | 5 | 10 |
| Result retention | 7 days | 90 days | 1 year |
| Start free | Subscribe | Subscribe |
Top-up packs
For occasional spikes beyond your plan’s monthly allocation. Top-ups don’t expire as long as your subscription is active. 500k+ packs require a Quant subscription.
| Credits | Price | Effective rate | Available to |
|---|---|---|---|
| 20,000 | $5 | $0.25/1k | All paying plans |
| 100,000 | $20 | $0.20/1k | All paying plans |
| 500,000 | $90 | $0.18/1k | Quant only |
| 2,000,000 | $300 | $0.15/1k | Quant only |
Refund policy
7-day money-back ONLY for first-month users with under 50% credits used. Beyond day 7 — no refunds. Subscriptions run until current period_end on cancel. Infra-error charges are auto-refunded.
What's a credit?
A credit is a unit of data processed, not a unit of time. One credit = 10,000 bars in minute mode (≈7 days of crypto 1-min data). A 1-year BTC minute backtest uses ~53 credits. Hybrid mode is 3× cost; tick is 10×. Wall-clock time varies with cold-start and mode — credits do not. Failed jobs are not charged. The MCP response includes your remaining balance.
Free-tier rolling refresh
Each free user gets 500 credits at signup. The clock starts then — 30 days later, balance is topped up to MAX(current, 500). No stacking, no rollover beyond plan limit.
Fair-use clause
Sustained usage above 3× your plan’s monthly allocation may be throttled or invoiced separately. We’ll contact you first — this is a safeguard against abuse, not a gotcha for power users.
Why credits, not a flat sub?
Different strategies cost very different compute. A 5-year tick backtest is 50× a 3-month minute backtest. Credits make this transparent — you pay roughly what we pay to run it, with a sustainable margin we’re honest about.
Why no enterprise tier?
We’re support-light by design. Heavy investment in docs and FAQ instead. Re-evaluating at 250+ paying users.
Still have questions? Read the FAQ →