01
A live portfolio twin
An accurate, up-to-the-minute picture of everything you hold — read from your broker and preserved as a point-in-time record.
Varyntra is an autonomous risk desk for your portfolio — watching every position around the clock, moving early to defend your downside, and held within limits it can never override.
The intelligence proposes.
Policy controls execution.
Varyntra's models read your portfolio and reason about risk. Acting on that — if it ever happens — takes a separate, authorized, human-controlled step the intelligence has no access to. Advice is powerful; authority is earned, and kept out of reach.
Nine disciplines, from a live view of your holdings to years of back-testing — each one showing its work, and each one answerable to the same set of limits.
An accurate, up-to-the-minute picture of everything you hold — read from your broker and preserved as a point-in-time record.
How much you could lose, and where the risk concentrates — calculated and repeatable, never estimated by a model.
Practical hedges — protective puts, spreads and collars — with the cost and payoff of each laid out plainly. Proposals, never orders.
Each option is weighed against your budget and limits, then resolved plainly — hold, or protect — with every reason on the record.
Plain-language commentary, grounded in the real figures. It can read everything about your portfolio, and change nothing.
Continuous observation as conditions shift, surfacing only what genuinely matters — so you are informed, not flooded.
Every intervention is measured against a do-nothing baseline, so the protection's worth is shown in dollars, not claims.
Strategies replayed across years of market history — and candid about the limits of what those results can claim.
An independent stop, tamper-evident approvals and safe-by-default settings — so the limits hold even under stress.
Link your account and Varyntra assembles a live, accurate picture of your holdings. Read-only to begin — it can observe, not touch.
Around the clock, it measures your risk, catches trouble early, and prepares practical ways to protect your gains — reasoning shown.
Every protective move waits on your approval. The intelligence advises; anything acted on happens only when you decide — on a paper account.
Every idea travels the same path, in the same order. Watch it stop: nothing reaches your broker without a separate, deliberate approval — and a switch that only you control.
The intelligence cannot directly reach your broker — and neither can a shortcut. Governed autonomous execution is possible only after deterministic approval and every execution gate permits it.
Not marketing claims. Each is enforced by the software itself and backed by tests.
Every figure that matters comes from a tested calculation — never from a model's guesswork.
Snapshots and decisions are locked once made, so the record always reflects what truly happened.
Your brokerage sits behind a strict boundary, and real-money trading is turned away outright.
Anything the AI says is checked against the real figures before it ever reaches you.
The do-nothing baseline can't be quietly reset, and gains are never counted twice.
Historical tests use clearly labeled estimates — never fabricated trades dressed up as real ones.
Open the Control Room to inspect the connected Alpaca paper account. Autonomous workers can trade only when every execution gate is deliberately released.
Alpaca paper endpoint only · Not validated against a real-money account · Not investment advice.
Inspect Varyntra's connected Alpaca paper account and its governed execution state.
No account yet? Open the live demo →
Concept preview · sign-in is illustrative and opens the Control Room.