Every plan runs the same models, the same regime engine, the same five intelligence services. Nothing is held back for a higher tier. What a subscription buys is depth of history and capacity to run new work — and you can calculate the second one yourself, because we publish the formula.
Same platform. Different capacity.
Billed monthly. Cancel any time.
Every model. Every service. The same engine.
5 years of history and 100 credits a month.
One exception, stated rather than buried: strategy search is a Professional action. The smallest search worth running is 500 walk-forward trials at 100 QC — the whole of this plan's 100 QC month.
For a self-directed investor with a process, working a focused book and checking a view they have already formed against independent evidence.
Billed monthly. Cancel any time.
Every model. Every service. The same engine.
The full archive, 1,000 credits a month, rolling over to 2,000.
For advisers, family offices and quantitative researchers who run standing analysis across several books and test their own policy against history.
Priced on seats, credit pool and data needs.
Every model. Every service. The same engine.
The full archive and 15,000 credits a month, pooled across the team.
For desks that need a shared pool, dedicated capacity, programmatic access and an audit trail behind every number a client sees.
| Entitlement | Investor | Professional | Institutional |
|---|---|---|---|
| Quant Credits | 100 QC / month | 1,000 QC / month | 15,000 QC / month, pooled |
| History depth | 5 years | Full archive | Full + custom ingest |
| Rollover | — | Up to 2,000 QC | Pooled, 12-month expiry |
| Overdraft buffer | — | 100 QC | Agreed in contract |
| Compute queue | Standard | Priority | Dedicated GPU |
| Strategy search | — | Included | Included |
| Symbols researchable | Unlimited | Unlimited | Unlimited |
| Models and services | Full | Full | Full |
| Standing daily analysis | 1 book | 10 books | Unlimited |
| Team workspaces, shared pool | — | — | Included |
| REST / WebSocket API | — | — | Included |
| Single sign-on, audit log | — | — | Included |
| Named quantitative engineer | — | — | Included |
Every tier runs the same models and the same actions, with a single exception we would rather state than bury: strategy search is a Professional action. It costs 20 credits per 100 walk-forward trials, so the smallest search worth running — 500 trials — is 100 credits, which is the whole of an Investor month. Offering it there would mean spending a month's capacity on one run and finding out afterwards. Everything else is limited only by credits and history depth, both of which map to a real cost and both of which you can work out in advance.
Investor reads five years; the full archive is Professional. This gates nothing functionally — every action still runs — but it maps to genuine storage and streaming cost, it is one sentence to explain, and it makes the upgrade a capability you discover you want rather than a wall you hit.
| At the monthly allowance | Investor | Professional | Institutional |
|---|---|---|---|
| Forecast refreshes | 25 | 250 | 3,750 |
| Single-symbol backtests, at the plan's full depth | 16 | 111 | 1,666 |
| 50-asset optimisations, one-year covariance | 4 | 41 | 625 |
| 50-asset full-history backtests | — | 37 | 555 |
| 1,000-trial strategy searches | — | 5 | 75 |
Professional is deliberately over-provisioned. Someone who runs one portfolio optimisation every working day and refreshes fifty forecasts a month consumes roughly 700 of 1,000 credits. Credits are meant to govern the extreme, not tax the ordinary.
Work out your month before you commit.
Set the work you expect to run. The estimator applies the same function published above — base rate × breadth × depth — subtracts each plan's allowance and prices any shortfall at the best pack rate.
| Action | Runs a month | Assets | History | Each | Line |
|---|---|---|---|---|---|
| Forecast refreshes | Current state | 44 × 1 × 1 | 200 | ||
| Intrinsic-value re-estimates | Current state | 66 × 1 × 1 | 30 | ||
| Backtests | 123 × 2 × 2 | 96 | |||
| Portfolio optimisations | 248 × 3 × 1 | 504 | |||
| Strategy search | 1 | Current state | 2020 × 1 × 1 | 40 |
this month includes a strategy search, which is not offered on this plan
1,000 credits included — this month fits inside them
the pooled allowance covers this several times over
Packs are one-off purchases, so a quieter month costs only the subscription — and on Professional an unspent grant rolls over. The 100-credit overdraft buffer covers a single overshoot mid-analysis; it is not a way to run a standing deficit.
Buy capacity when a research week runs hot.
Packs are one-off purchases, valid twelve months from the date you buy them, on any tier. They are consumed only after the monthly allowance is exhausted, and they survive both downgrade and cancellation.
| Pack | Price | Per credit | Position |
|---|---|---|---|
| 250 QC | $30 | $0.120 | Occasional overflow |
| 1,000 QC | $100 | $0.100 | A second month of Professional compute |
| 5,000 QC | $425 | $0.085 | Sustained parameter studies |
Reaching Professional's allowance from Investor costs $120, against $80 for Professional — which also carries the full archive, rollover, the priority queue and the overdraft buffer. Reaching Institutional's pooled allowance from Professional costs $1,330, against $1,200. This property is re-verified on every price change; it is the single most common way credit ladders break.
Unused Professional credits roll forward to a ceiling of 2,000. Uncapped rollover is an open-ended liability — a dormant account banks a year of capacity and can release it in a week — and the cap holds that exposure down while still rewarding a quiet month.
You are never blocked mid-analysis. Going negative is permitted up to the buffer and settles against next month's grant automatically. It is there so that the worst experience a metered product can produce — a run that dies halfway — cannot happen.
Stated plainly, with the reason: it keeps compute demand predictable and stops a dormant account from banking capacity it releases all at once. Purchased packs are unaffected, and Investor has no overdraft buffer because it has no rolling balance to settle against.
Rolled monthly credits are forfeited; purchased pack balances are not, for the full 12 months. Nothing is ever auto-charged to cover a shortfall — if a month runs past its allowance the run waits for you to confirm a pack.
Before you subscribe.
No. Professional accounts carry a 100-credit overdraft buffer, so a run that takes you past zero completes and settles against next month's grant. You are never stopped in the middle of a piece of work, and nothing is charged without your confirmation.
Every priced action quotes itself first, showing the base rate and the multipliers applied. You confirm, then it runs. The formula is published above, so you can also work it out yourself.
No. Symbol caps were removed. Reading research on any company, on any plan, is free and unlimited. Credits apply only when you ask the platform to produce something new.
On Professional they roll over to a ceiling of 2,000. On Investor the monthly grant expires at cycle end, which keeps compute demand predictable. Purchased packs are separate: valid 12 months, on any plan, surviving both downgrade and cancellation.
Yes. Set a monthly ceiling in account settings and the platform stops rather than bills. Off by default.
Both directions, prorated. Upgrades take effect immediately with the new allowance available the same day. Downgrades take effect at renewal, so you keep what you have paid for.
No, and it is the only action any plan withholds. A search is priced per 100 walk-forward trials, so the smallest one worth running — 500 trials — costs 100 QC, which is an entire Investor month. Offering it there would mean spending a month's capacity on a single run and discovering it afterwards, so we say so instead. Every other action runs on every plan.
30 days of every Professional feature, and a credit grant of 0. The zero is deliberate rather than an oversight: credits are compute, compute costs us on every run, and bundling it into an evaluation window prices the evaluation instead of the product. Reading research is free and unlimited throughout, exactly as it is on a paid plan; if you want to run priced work before you subscribe, a top-up pack does it. No card to begin, and it does not convert unless you choose a plan.