How it works

One command. One process. One screen.

myTbot is built around a single command. When the operator runs it, the platform brings up every subsystem in order, and only the subsystems whose credentials are present. Failure at any stage is loud and explicit.

bash — user@mytbot

$python run.py

Runtime sequence

01Dependency managerPins versions, verifies the Python and model runtimes are intact.

02Broker discoveryDetects which broker connectors hold valid credentials right now.

03Data pipelineStreams quotes, news, macro feeds. Stale data is dropped before it can influence sizing.

04Trading loopStrategy families generate intents on each cycle. Intents are graded, not executed.

05Risk engineVetoes or sizes the intent. The final authority before any order leaves the process.

06Execution routerSelects the venue with the best blend of fee, slippage and fill probability.

07FastAPI + WebSocketPublishes state to the dashboard. The operator sees what the system sees.

08React dashboardOne screen. A capital slider. A kill switch. Panels that explain every recent action.

The operator's surface

The dashboard is intentionally narrow. The operator does not pick trades. The operator configures boundaries, watches the system reason inside them, and stops it if needed.

/ Capital allocation slider
/ Kill switch
/ Live signal and intent panels
/ Risk engine state and recent vetoes
/ Fills ledger with decision heritage