Reviewing a parsed strategy

After submitting a strategy description, the interpret screen shows the parsed conditions, market config, and risk parameters — review all of them carefully before launching to ensure the AI understood your intent.

What is the interpret screen?

After you submit a strategy description, the interpret screen shows:

• The parsed entry and exit conditions (human-readable summary)
• Market configuration (symbol, venue, interval, asset type)
• Position sizing and risk parameters (stop loss, take profit, trailing stop)
• max_open_positions setting
• A link to view the generated Python code

Review each section carefully before launching.

How do you check entry and exit conditions?

The entry and exit conditions are shown as plain-English summaries derived from the generated Python. Read them to confirm the AI understood your intent. If they don't match, re-describe your strategy and re-submit — each submission is a fresh parse.

How does max_open_positions work?

This caps concurrent open positions:

• null — no cap (opens a new position for every signal)
• A positive integer — maximum concurrent positions
• 0 — blocks all new entries

How do you view the generated Python?

Click the code icon (or "View code") on the interpret screen to see the full Python source generated by the AI. The strategy subclasses Strategy from coinrule_x_strategies.core and implements indicators() and evaluate().

Review the code to verify:

• The indicators match what you described
• The entry and exit logic is correct
• No unexpected side effects

How does draft persistence work?

Your parsed strategy is saved in browser session storage after a successful parse. It is lost if you close the tab before launching. There is no server-side draft — submit a new parse if you lose the session.

How do you launch a strategy?

Click Launch to deploy the strategy. At this point:

1. The strategy is saved to the database
2. The engine assigns it to a runner and begins watching the market
3. A background task generates the test file (using Sonnet) — this happens automatically

The strategy appears in your agents list immediately with a "Deploying" status, then transitions to "Running".