NexusTrade Financial MCP

NexusTrade TypeScript SDK

Author trading strategies in typed TypeScript. Backtest them on the engine that runs them live.

Quickstart · Authoring · Polling · Agents · Lake SQL · Auth · Errors

---

npm install nexustrade

Zero runtime dependencies. ESM and CommonJS builds ship together, with types.

MCP server

NexusTrade also exposes the platform as a hosted, remote Model Context Protocol server. Modern MCP clients connect directly to the production Streamable HTTP endpoint and discover NexusTrade OAuth automatically:

claude mcp add --transport http nexustrade https://nexustrade.io/api/mcp

Cursor and other remote-capable clients use:

{

"mcpServers": { "nexustrade": { "url": "https://nexustrade.io/api/mcp" } } }

For Claude Desktop and other stdio-only clients, use the established mcp-remote bridge—no clone or local NexusTrade server is required:

{

"mcpServers": { "nexustrade": { "command": "npx", "args": [ "-y", "mcp-remote@latest", "https://nexustrade.io/api/mcp", "--transport", "http-only" ] } } }

The live server exposes more than 120 tools across market research, portfolio construction, backtesting, optimization, walk-forward validation, managed compute, Aurora agents, paper trading, and controlled brokerage operations. Its creator-marketplace tools cover the full strategy adoption path:

• search_creators discovers public creators and their marketplace portfolios.
• subscribe_portfolio validates a monetized listing and returns a safe checkout preview; the user completes payment in NexusTrade, never through the MCP tool.
• fork_shared_portfolio creates a one-time editable copy of a marketplace strategy in a new or existing portfolio.
• copy_trade_shared continuously mirrors an accessible strategy into a paper or live portfolio at an explicit allocation.

See the developer guide, the utility tool reference, and the Aurora tool reference.

Research and historical results are not investment advice and do not guarantee future performance. Keep paper and live modes explicit. Tools that can affect portfolios, schedules, or brokerage orders remain subject to the authenticated account's NexusTrade permissions and approval controls.

Quickstart

import {

NexusTradeClient, always, backtest, buy, portfolio, stockAsset, strategy, } from "nexustrade";

const client = new NexusTradeClient();

const book = portfolio("Example", [ strategy("Buy SPY", always(), buy(stockAsset("SPY"), 100)), ]);

const operation = await client.createBacktest( backtest(book, { startDate: "2024-01-01", endDate: "2024-12-31" }), { idempotencyKey: "example-v1" } ); const result = await client.waitForBacktest(operation.id as string); console.log(result.result);

Backtest operations may include warnings: string[] immediately after submission and again in the terminal result. Treat them as material caveats; they do not change a successful operation into a failure.

Collateral at risk

A terminal operation's result.statistics answers how much capital the run had on the line, not only what it returned. Two fields carry it, typed as BacktestCollateralStatistics:

Both are optional and may be null, and that absence is a real answer: a backtest run before the engine reported collateral has no value, which is not the same as a book that locked nothing. Display "not recorded" rather than $0 — a zero here reads as "this strategy risks nothing", the opposite of what an unpopulated field means. Do not substitute the portfolio's value either: a book risking a few thousand dollars would be reported as risking all of it.

Never reconstruct either number from cash - buyingPower. Buying power is clamped at both ends and carries an open credit-spread premium term, so the inversion breaks precisely on the heavily collateralised books this measures.

Authoring strategies

Every builder is generated from the same indicator specification the NexusTrade engine runs, so a book is valid by construction rather than by convention.

TypeScript cannot overload comparison operators, so indicators compose through gt / gte / lt / lte / eq / neq and and / or:

import * as nt from "nexustrade";

const book = nt.portfolio( "Momentum", [ nt.strategy( "Rotate into strength", nt.always(), nt.dynamicRebalance({ universe: nt.universe("SP500"), pipeline: [ nt.filter(nt.gt(nt.Price(nt.CANDIDATE), nt.SMA(nt.CANDIDATE, 200))), nt.selectTop(nt.RSI(nt.CANDIDATE, 14), 10), ], weightIndicator: nt.RSI(nt.CANDIDATE, 14), limit: 10, deploymentPercent: nt.Value(80), }) ), ], { initialValue: 100_000 } );

Congressional-disclosure signals are native indicators too. House and Senate are combined unless you explicitly choose a chamber; amount metrics default to the disclosed lower bound and confirmed option disclosures are excluded by default:

const pelosiPurchases = nt.PoliticalTrades(

nt.CANDIDATE, "Nancy Pelosi", "BuyAmount", 90, "LowerBound", "Equity", );

Use "" as the filer to include all members. BuyCount counts canonical purchase events, while BuyAmount measures disclosed magnitude. Select "Option" only when you want option-disclosure activity associated with the underlying ticker.

Order execution belongs to the strategy. Omit it for the backward-compatible Market default, use a fixed unit price for Buy/Sell, or set an option strategy's maximum net debit / minimum net credit:

nt.strategy("Buy SPY at my price", nt.always(), nt.buy(nt.stockAsset("SPY"), 10), {

orderExecution: nt.limitOrder({ price: nt.unitPriceLimit(500), workingTime: nt.goodForDay(), }), });

nt.strategy("Sell the spread for $1.50 or better", nt.always(), optionAction, { orderExecution: nt.limitOrder({ price: nt.minimumNetCredit(1.5), workingTime: nt.goodForMinutes(30), }), });

An option net limit can also follow a position the strategy holds. Pass an indicator instead of a number and the engine evaluates it when the strategy fires. If it has no positive value at that moment, no order is placed:

nt.strategy("Sell the credit spread for the debit fill plus $0.50", whenDebitSpreadHeld, creditSpreadAction, {

orderExecution: nt.limitOrder({ price: nt.minimumNetCredit( nt.Plus(nt.OptionSpreadEntryPrice("SPX", "call", "long", "vertical"), nt.Value(0.5)), ), workingTime: nt.goodForDay(), }), });

currentLimit() creates a quote-relative Limit for dynamic rebalance strategies. It keeps the no-worse-than-current-quote protection, but it is not a resting price target. Live option strategies must choose an explicit Limit policy.

Option selection and execution limits

strategy(..., { pendingOrderPolicy: "CancelConflictingBuys" }) enables backtest-only exit priority over conflicting pending buys. Omission preserves waiting. Paper and live deployment reject this opt-in because broker cancellation acknowledgements are not implemented. The setting also accepts explicit "Wait".

Builders describe drafts. Typechecking does not prove that the engine can execute every requested rule; the API validates complete portfolios and their conditions.

• Expiration selectors support a DTE window and nearest, middle, or furthest. middle selects the middle eligible expiration by index, not the expiration nearest a target such as 180 days. There is no monthly-only or quarterly-only expiration filter.
• rebalanceOption accepts renewBeforeDays (integer 1–90) and revisitDisclosedTargets. These maintain modeled political-disclosure option targets. The strategy must contain a NewPoliticalDisclosure condition for instrument Option with exactly one member ID. These fields do not implement a general close-and-replace roll for a trend strategy.
• For backtests, the job's feeConfig.OptionSlippageFraction sets quote-side fills: 0 uses the midpoint, 1 uses ask for buys and bid for sells, and the default is 0.5. This job setting does not add a last-trade fallback or define live order execution.
• Evaluation timing and order execution are separate. OpenClose schedules evaluations at session open and close; it does not itself carry a signal observed at close into an order at the next open. Explicit zero timing offsets and a gt(MinutesAfterOpen(), Value(0)) condition can restrict evaluations to the calendar session close. MinutesUntilClose counts down to 16:00 Eastern, so it cannot identify an early session close.

When authoring an approximation, describe the actual expiration policy, signal timing, fills, and replacement behavior. Backtest results describe that implementation, not the exact method it approximates.

What you can build — 170+ generated builders

Every builder is fully typed — your editor completes the whole surface.

Jobs run on the engine — you poll

create* enqueues work and returns immediately. It does not resolve when results exist. There are no webhooks today.

sequenceDiagram

participant You participant SDK participant Engine

You->>SDK: cr