Every job a trading day demands

Observe → Reason → Act → Report, around the clock. Entries are conviction-gated: no conviction, no trade — and every decision is journaled with its reasoning, including the trades it passed on.

Built-in paper accounts fund and fill locally, or connect venue-native paper at Alpaca. Promote a setup to live capital on Alpaca or Kraken only once its journal has convinced you.

Stop-loss and take-profit brackets (OCO) travel with the signal from the moment it's placed — venue-native where the exchange supports it, engine-managed where it doesn't.

An editable commission and fee schedule is the single source of truth — the same table charges your paper fills, estimates your live costs, and prices your simulations. No fee-free fantasy P&L.

Run strategies against real market products with realistic fees, grouped by product and comparable side by side — so you know what works before money moves.

Describe what you want to test and a planner pre-flights the run; an agent trader then manages the simulation live, switching strategy mid-run when the market stops cooperating.

One click renders a share card from a finished simulation and publishes it to the Londron network — and on to Instagram — receipts included.

Tell it what to watch — a level, a pattern, a condition. Sightings route automatically to an agent, a board, or a log, so the desk reacts to the market instead of polling it.

A human-in-the-loop gate enforced by the engine, not by prompt politeness: automated routes can propose actions, but anything consequential waits in your inbox for a yes.

Recurring agent tasks on your cadence — a pre-market scan at 07:00, a weekend deep-dive, a Sunday portfolio review. Every run keeps its own log.

Economic events and earnings stream in and get ranked by an AI triage pass. The ones that matter arrive pinned and annotated before they hit the tape.

Run a local model on your own hardware or connect a cloud API — per module. “Test & auto-configure” probes the model and applies working settings for you.

Ask in plain language. Answers come from live quotes and the agent's own journal — every figure traceable to a tool call, never a guess.

Screening, research synthesis, calendar triage, news classification — each agent runs on its own prompt and its own model configuration.

Token usage is tracked per module and per provider, so you always know what your intelligence actually costs.

Expose London5's tool registry to external agents over MCP — loopback-only, bearer-authenticated, and permission-tiered. Your desk becomes a tool your other agents can call.

Overnight activity condensed into a readable report: what it saw, what it did, what it wants you to look at. Ready when you wake up.

A movable widget grid: performance hero, positions, market strips, clocks, and more — arrange the cockpit the way you actually work.

Quotes stream through a single cached quote layer to every surface — no refresh buttons, no stale numbers, and symbols follow one convention across the whole app.

Per-market symbol directories cover the complete US equity and ETF universe — plus a wheel-of-fortune picker for when you want the market to surprise you.

Session clocks show what's open where, right now; market-sentiment gauges by asset class feed the same context the agent trades on.

Engine, data, and decisions live on your device. There is no London5 server in the trading loop.

Broker and exchange keys sit in your system keychain and speak only to official venue APIs.

Every decision recorded with its full reasoning — auditable, reviewable, yours.

Position caps, loss limits, and exposure ceilings enforced by the engine — and one click flattens everything.