Guides / Getting started

Getting started · about 5 min

Your first decision in the Playground

Ask three trading questions about a stock with nothing but its symbol, read the probabilities, and turn one of them into a decision. No code and no API key.

1. Open the Playground

   Click Playground in the left menu. It has three inputs on the left (State, Questions, Actions) and the response on the right. Runs are billed like API calls and show under Usage as “Playground”.

2. Load a request

   The state is just the symbol: {"symbol": "TSLA"}. The server fills in TSLA's latest completed session (returns, volatility, distance from highs and moving averages, volume, earnings timing). The questions are three yes_no forecasts, each with a clear outcome and a horizon:

   {
     "model": "ora1",
     "state": {
       "symbol": "TSLA"
     },
     "questions": {
       "drop5": {
         "type": "yes_no",
         "instructions": "Will it drop at least 5% at some point in the next month?"
       },
       "rally5": {
         "type": "yes_no",
         "instructions": "Will it rally at least 5% at some point in the next month?"
       },
       "up_week": {
         "type": "yes_no",
         "instructions": "Will the stock be higher in a week?"
       }
     }
   }

   Open this in the Playground

3. Run it

   Press Run (or Ctrl/⌘ + Enter). A real answer, for the session of 2026-10-30:

   {
     "model": "ora1",
     "version": "ora1-fin-0.1.1",
     "answers": {
       "drop5": {
         "type": "yes_no",
         "yes_no": 0.638
       },
       "rally5": {
         "type": "yes_no",
         "yes_no": 0.687
       },
       "up_week": {
         "type": "yes_no",
         "yes_no": 0.487
       }
     },
     "usage": {
       "input_tokens": 363,
       "output_tokens": 12
     },
     "context": {
       "source": "server",
       "symbol": "TSLA",
       "as_of": "2026-10-30"
     }
   }

   context tells you which session the answer is based on. Your numbers will differ: they move with every session.

4. Read a yes_no probability

   A yes_no answer is the probability that the outcome happens, from 0 to 1. Read each one as odds, not as a verdict:

   • drop5 0.638: about a two-in-three chance that TSLA trades at least 5% lower at some point in the next month.
   • rally5 0.687: a similar chance of a 5% rally in the same month. Both being likely means a wide range, not a direction: two-way risk.
   • up_week 0.487: close to a coin flip, so no edge on where it closes a week from now.

   Values near 0.5 carry little information; the further from the outcome's usual frequency, the stronger the signal. Read probabilities and choose thresholds shows how to pick the line.

5. Turn it into a decision

   Answers are numbers, so a decision is a comparison. For example, cut position size when the drawdown odds are high, and only act on direction when it is clearly away from 0.5:

   a = data["answers"]
   if a["drop5"]["yes_no"] >= 0.6 and a["rally5"]["yes_no"] >= 0.6:
       size_down("TSLA")          # wide range both ways
   elif a["up_week"]["yes_no"] >= 0.6:
       add_to_watchlist("TSLA")   # a directional lean

   To have the rule run on our side and post to Slack or your own system, add it as an action: Send a Slack alert when a rule fires.

6. Keep it

   Press Save to keep the request under Saved queries, or Code to get it as cURL, Python, JavaScript and more.

7. Optional: ask about text

   The same request shape works for any text or record: put the content in state and ask a choice (pick one option) or a score (rate on ordered levels). Try it with a support ticket:

   Open the support-ticket example

What's next

• Read probabilities and choose thresholds
• Call the API from your code
• Send a Slack alert when a rule fires