Skip to content

Session Architecture

In Cortiq, the session is the root of the trading architecture.

If you understand the session correctly, the rest of the trading-cycle entities make sense much faster.

A session controls four things at the same time:

  1. Trading scope
  2. Strategy scope
  3. Execution scope
  4. Support context

That means one session can define a very different operating model from another, even when both use the same broker terminal.

Session As The Customer-Facing Architecture Unit

Section titled “Session As The Customer-Facing Architecture Unit”

When a customer creates or edits a session, they are really assembling a trading system from reusable parts.

Session AreaWhat The User Chooses
Broker and market scopeMT5 account, fixed symbol or AutoScan behavior
AI routingProvider, integration mode, and fallback behavior
StrategyPlaybooks and their priority
DataData package and indicator scope
Support contextPreparation package, instrument profile, sentiment report, and instructions
Operating windowActive days, hours, and close-before-end behavior
Execution modeLive, virtual, copy-trading behavior, and notifications

Why The Session Matters More Than A Single Prompt

Section titled “Why The Session Matters More Than A Single Prompt”

The session is not just a saved configuration form.

It is the thing that makes the trading cycle repeatable.

Instead of rebuilding the whole environment every time, the user creates a session once and then:

  • starts it
  • pauses it
  • reviews it
  • improves it
  • duplicates it into a new operating template

At runtime, the session sits in the center of the loop.

Session -> Data Gather -> Prompt Build -> AI Decision -> Execution -> Timeline

That is why customers should think of sessions as operating containers, not just presets.

Support in this context means any information the AI can use beyond raw current candles.

The session can add:

  • playbooks for rule-based setup logic
  • trade ideas for specific active theses
  • instructions for operator guidance
  • preparation outputs for slower-moving structure
  • an instrument profile for symbol behavior
  • a sentiment report for macro and headline context

Two sessions can trade the same symbol and still behave very differently.

Example:

  • Session A uses a narrow London breakout playbook, a lean data package, and no support layers.
  • Session B uses a swing-oriented playbook stack, a broader multi-timeframe data package, a preparation package, an instrument profile, and a fresh sentiment report.

Even on the same symbol, those are not the same operating system.

Read Supporting Context to see what types of information can be attached to a session and when each one is useful.