Write a portfolio rule as a sentence. Compile it into a Script with explicit triggers, sizes and guards. Simulate it against history before anything runs — across equities, RWAs, crypto and commodities from one execution layer.
The app is wallet-gated. The demo needs no wallet.
A strategy is not a chart or a hunch — it is a rule with a threshold, a size, a source and a limit. Move the controls and the backtest re-runs the same trigger and guard code the runtime uses.
Demonstration dataset · benchmark: static 30/70 BTC / Treasury · 0 triggers blocked by guards.
Illustrative figures. Simulations inside the product run on retrieved market history and are labelled with their source.
Written down, a rule can be tested. Tested, it can be trusted. Trusted, it can be automated. Everything between those three states is where portfolios quietly drift away from the intentions that created them.
flowta makes the written form the primary object: a Script you can read, simulate, argue with, fork and constrain — across equities, crypto, RWAs and commodities at once, rather than one venue at a time.
Portfolio intentions become structured strategies with explicit triggers, actions and guards.
Understand behaviour across a real window, against a benchmark you choose, before activation.
Coordinate equities, crypto, RWAs and commodities from a single strategy layer.
Exposure ceilings, reserve floors and cooldowns define what the runtime may never do.
Five stages, in order, each one inspectable. Nothing about a Script is hidden from the person who wrote it.
Type what you want your capital to do, in plain English. The compiler runs locally in your browser — no model call, nothing leaves the page.
“If BTC falls 15% from its 30-day high, allocate $1,000 from my Treasury reserve.”
Intent becomes a versioned, schema-validated definition. Confidence is reported and every assumption is stated — the compiler will refuse before it will invent.
The backtest steps daily through retrieved market data and runs the same trigger and guard code the live runtime uses. Sourced history and demonstration data are never blended silently.
Exposure ceilings, reserve floors, transaction caps and cooldowns are evaluated before any action is applied. A blocked trigger is recorded, never silently skipped.
Activation begins the monitoring loop, computes the next evaluation and writes activity. In this release the runtime simulates — no order reaches a venue.
The same definition drives the composer, the backtest and the runtime — so what you read is exactly what runs.
The condition that wakes the Script. Price levels, percentage moves, drawdown from a rolling high, schedules, allocation drift or a yield threshold.
The capital movement it proposes. Sized in dollars or as a share of the position or portfolio, always with an explicit source and destination.
The limits the runtime may not cross. Evaluated before the action is applied, so a Script that would breach a limit is stopped and the block is recorded.
A Script can reference an equity, a token, a tokenized Treasury and a metal in the same rule. Every entry carries its own provider, chain and freshness.
| Asset class | Count | Registry entries | Freshness | Source |
|---|---|---|---|---|
| Equities | 09 | NVDAAAPLMSFTMETATSLAAMZNGOOGLSPYQQQ |
Delayed | Market data vendor |
| Crypto | 05 | BTCETHSOLLINKAVAX |
Live | Public market data |
| RWAs | 06 | USTBUIDLBENJIUSTBUSDYPCRED |
Registry | Issuer + chain |
| Commodities | 06 | XAUXAGWTIBRENTNATGASCOPPER |
Delayed | Market data vendor |
| Stable | 03 | USDCUSDTDAI |
Live | Onchain |
The last figure is the important one. flowta simulates: Scripts are evaluated, guarded and logged, but no order reaches a venue and no capital moves.
Curated Scripts you can read in full, simulate on your own window, then fork into your drafts and change every field.
Accumulates BTC during large drawdowns while maintaining a stable reserve floor and a hard ceiling on crypto exposure.
Realises a portion of ETH gains into Treasury exposure whenever the position advances beyond a defined threshold.
Adds to concentrated large-cap technology exposure on confirmed monthly strength, with a strict single-name ceiling.
Figures are simulations over the stated window, produced by the flowta backtest engine. They are not records of realised performance and no return is guaranteed.
Constraints are easier to trust than promises. These are the ones the architecture enforces.
flowta never holds your assets and cannot move them. Connecting a wallet proves you control an address — it grants no spending permission and no signing authority.
No email, password, social login or passkey. One authentication path, and a wallet address supplied in a request body is never treated as proof of identity.
Row level security is on for every table. Identity-bearing operations run through one function that verifies your session token server-side and derives the account from the verified payload.
If a provider cannot supply a value, the interface shows an em dash and says why. Live, delayed, cached and demonstration values are visually distinct everywhere.
No account, and no password. flowta authenticates with a wallet — connect an EVM wallet and your Scripts, portfolio and activity bind to that address. If you want to look around first, the demo runs every screen with no wallet at all.
No. flowta is non-custodial. It never holds funds, cannot sign a transaction, and in the current release does not transact at all — the only execution adapter is the simulation adapter.
No. The default compiler is deterministic and runs entirely in your browser. There is no model call, no API key, and nothing about your intent leaves the page. The architecture allows a model-backed compiler later, behind the same interface.
Where a provider supplies history, yes — and the report is labelled Live history. Where it does not, the engine falls back to a bundled dataset and labels the report Demonstration dataset. The two are never blended silently.
No. flowta provides simulation and portfolio automation previews and does not constitute investment advice. Nothing in the product recommends what to buy or sell, and no return is guaranteed.
2 Scripts monitoring markets in the simulation runtime.
Rules currently evaluated against market and portfolio state
Net value by asset class
Observations from current data — not advice
One registry across equities, crypto, RWAs and commodities.
| Asset | Price | 24H | 7D | Trend | Market / type | Volume | Status |
|---|
Positions, allocation and risk across every asset class.
Marked to the latest resolved price for each asset
| Asset | Quantity | Price | Value | 24H | PNL | Weight |
|---|
Analytics derived from current positions — not recommendations
Rules that move capital.
Backtest any Script against retrieved market history, then stress the resulting book.
Deterministic engine — the same trigger and guard code the runtime uses
Pick a Script and a window, then run the backtest to see performance, risk and the execution timeline.
A linear shock applied to current exposures
A scenario estimate, not a prediction. Shocks are applied linearly to current exposures; correlations, liquidity and path dependence are not modelled. Rate moves apply a 1.8-year duration proxy to the RWA leg.
Discover strategies built for autonomous markets.
Figures are simulations over the stated window, produced by the flowta backtest engine. They are not records of realised performance and no return is guaranteed.
Every runtime event, in one log.
flowta's current MVP provides simulation and portfolio automation previews and does not constitute investment advice. No order reaches a venue and no capital moves.