Playground
The Playground is a page in the console where you ask questions about your own account in plain language — your lists, groups, models, media, services and search analytics — and an AI agent looks the answer up for you and lays it out properly: real tables you can sort, cards, key figures. It’s the fastest way to explore what’s in your account without clicking through every page.
The playground is read-only. It can look anything up but it cannot create, change or delete anything, so there’s nothing to be careful about — ask it whatever you like.
Auto-plays · use Back / Next to step through at your own pace.
Asking questions
Open Playground in the console sidebar, type a question and press Ask. The answer streams in as it’s worked out, and under each answer a collapsible “n tool calls” trace shows exactly which look-ups the agent made to get it — the answer is always drawn from your live data, never guessed. Questions that suit it well:
- “Which of my lists have no model attached?”
- “How many records are in each of my lists?”
- “Which lists get the most searches, and what’s coming back empty?”
The answer is laid out to match what you asked for: a set of rows comes back as a real table you can sort by clicking a column heading, and key figures come back as cards.
An optional Focus selector under the question box points the agent at one list or group, so “how is it doing?” means that list. It’s a hint, not a wall — account-wide questions still work while focused. List pages carry a shortcut that opens the playground already focused on that list.
Weekly AI credits
Each answer costs real AI usage, so every account gets a weekly allowance of asks: 10 a week free, 20 a week on smaller paid plans, and 50 a week on larger ones. A quiet meter on the page shows where you stand, and the week resets on Monday. The allowance is shared with the assistant panel that appears on other console pages.
Bring your own Anthropic key
If the weekly allowance isn’t enough, add your own Anthropic API key from the Use your own key button: usage then bills to your own Anthropic account and the weekly meter no longer applies. The key is checked against Anthropic when you save it, stored encrypted, and never shown again; remove it whenever you like. You’ll need an Anthropic account — the dialog links to where keys are created.
On your own key you also get a Model selector, from fast and economical up to the most powerful models available — your key, your choice of horsepower.
Bigger jobs: Open in Claude
The playground is built for quick exploration, not for building things. Ask it for something genuinely big — “build me a web page from my product data”, a long report, code — and rather than half-doing it, the agent offers an Open in Claude button. That opens Claude with your task pre-filled and instructions to connect the Search Stack connector, so Claude can work with your live records at full power and build the real thing. Nothing is sent anywhere until you click it.
Privacy and safety
The agent acts as you: it can only read what your own sign-in can read, and it runs against your account alone. Answers render as data, not as live code — anything the model produces is treated as untrusted and displayed safely. And because the playground is read-only end to end, no question you ask can modify your records, your settings or your services.