An open source project from Emfasys Labs, released under AGPL v3
myTbot mascot — friendly trading robot

A personal autonomous trading system you run yourself.

myTbot is a free, open source desktop application. It installs on the operator's own machine, connects to their own broker account, and runs a transparent trading loop under a risk engine with final veto authority.

No accounts on our side. No custody of funds. No signal subscriptions. The operator holds the keys, the capital and the kill switch.

bash — user@mytbot
$python run.py --mode=paper --env=local
01 / What it is

A research platform, not a product

myTbot is infrastructure. It is not a signal service. It is not portfolio management. It does not hold capital. The operator is the only safety net.

02 / Why it exists

Transparency over claims

Most automated trading systems are closed boxes. myTbot publishes its decision heritage. Every behaviour traces to a specific recorded decision in the source tree.

03 / How it runs

One command, one process

python run.py starts the dependency manager, broker discovery, trading loop, data pipeline, API and dashboard. Nothing is hidden.

Architecture

Six layers, one direction of flow.

Each layer can run on deterministic logic if the AI layer is unavailable. AI never places orders.

Read the architecture →
01Discovery
Universe funnel across equities, ETFs, bonds, forex, crypto
02Conviction
Strategy signal plus AI scoring and news context
03Allocation
Sizing, displacement and cluster caps
04Execution
Venue selection, fee and slippage routing
05Containment
Risk engine veto, the final authority
06Audit
Fills ledger, decision heritage, deterministic logs
D125 · CONTAINMENT

The final authority

The risk engine holds veto power over every order.

Every trading intent passes through the containment layer before it reaches a venue. If a hard cap is breached, the order is refused. AI provides advice. The risk engine provides enforcement.

Read how the veto works →
Activation path

Paper, then micro-live, then live.

Phase 01

Paper

The default. The system runs end to end against live market data and never sends a real order. Use this to soak the loop for weeks.

Phase 02

Micro-live

Tiny capital, strict caps. The first real fills surface real venue behaviour: partial fills, rejections, slippage, authentication drops.

Phase 03

Live

Full operator capital, full risk engine authority, full audit. The operator stays the final safety net.

myTbot is research infrastructure. It is not investment advice. It is not a brokerage. It is not custody. It is not portfolio management. It is not a signal service. It makes no promise of performance. Trading carries the risk of capital loss.