GuidesStart free
Read the receipt under an answer
Every answer in Keenoble carries a badge showing which model answered, what it cost in Sparks, and why it was routed there. Here is how to read all of it.
Draft — not yet walked in the live product
All 7 steps below were written by reading the Keenoble source rather than by performing them, and each unwalked one is tagged [UNVERIFIED]. Reading a component tells you what a control is labelled; it does not prove the control is reachable, that the order is right, or that the result is what the text says. Until a reviewer has walked every step, this page is excluded from the sitemap, marked noindex, and publishes no structured data.
Before you start
- An answer on screen — see “Ask your first question in Keenoble”.
Steps
Look at the row underneath a finished answer. It holds, left to right: “Copy”, “Rate”, “↺ Regenerate” on the newest answer, the receipt badge, and “Total time”. [Unverified]
Read the receipt badge itself. It shows the model that answered, then a dot, then the cost in Sparks — for example a model name followed by “0.4sp”. [Unverified]
sp is Sparks, the unit the whole product is priced in. It is the same unit shown in the balance chip in the top bar.
Hover, focus or tap the badge to open the routing details. [Unverified]
It is a real button, so Tab reaches it and Enter or Space opens it. Escape closes it.
Read the routing lines. You will see one or more of: “You picked this model.”, “Routed automatically to fit the task.”, “Used a live data source, not the model's training data.”, or “Handled under stricter rules for a regulated topic.” [Unverified]
If the first-choice model was unavailable, the badge shows an arrow between two model names and the details say so: “First choice (…) was unavailable — answered by … instead.” [Unverified]
When a fallback changes the price, the cost is restated as “X → Y sp (corrected)” and the details add a line explaining the correction. The corrected number is the one you were charged.
Note the “Request …” line at the bottom of the details. Quote that id to support and we can find the exact request. [Unverified]
Read “Total time” next to the badge. That is the whole request — classification, any failed attempts, and the winning model's generation — not just the model's own time. [Unverified]
What a tool call looks like
When the router decides the honest answer is a live lookup rather than a language model, the badge names the tool and its provider instead of a model — a weather source, a market price source, a market news source.
If the tool was tried and failed, the badge says so explicitly (“… tool failed”) and the details add “Tried a live data source; it failed, so the model answered from what it knows.” You are never told a live source was used when it was not.
What the receipt deliberately does not show
The routing configuration. You get what happened — which model, why in plain words, what it cost — not the internal cluster ids, capability keys or candidate lists behind the decision.
That split is deliberate and it is the same one this lab applies to itself: findings and method are published, the current routing configuration is not.
Next
Having problems with the platform? Contact us at support@keenlabs.pro. If a step here does not match what Keenoble does, the step is the bug — send it to the same address. All the guides are at /guides; shorter answers are on the FAQ.