Please use our numbers. The measured queue times, percentiles, distributions and verdicts we publish are open data, licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). You may copy, republish, chart, quote and build on them — including commercially — for the price of a credit and a link back. That is the whole deal, and we would rather you used the data with credit than left it on the shelf.
We want these numbers quoted. Nobody else measures how long the batch APIs actually take, so every reuse makes the dataset more useful and every credit points a reader back to the source. The licence exists to make that use visible, not to restrict it.
What is covered
CC BY 4.0 covers the published aggregate measurements — everything on the open pages and the open API:
- the dashboard at
/and the per-model and per-provider pages, - the open JSON endpoints, keyless:
/v1/coverage,/v1/wait,/v1/curve,/v1/should-i-batchand/v1/estimate-batchtime, - the tables and figures rendered on
/compare,/how-we-measureand the other content pages, - the measured figures in
/llms.txtand/llms-full.txt.
These are aggregate statistics — percentiles, distributions and counts across many jobs. They are not, and never contain, an individual contributor's raw submission: the Terms §7 keep those private, and §8 lists what we never even receive.
What is not covered
CC BY 4.0 applies to the open, published aggregates. It does not cover paid or entitled data:
- Live, undelayed access and any feed under a paid plan is governed by the Terms, not by CC BY. A paying customer's live feed is not open data, and licensing it CC BY would let them republish the very thing they paid for.
- The site's own text, code, brand name and logo are not part of the data licence.
The line is simple: the published aggregates are CC BY 4.0; live, delayed-free access under a paid plan is governed by the Terms.
How to credit us
Attribution is the one thing we ask, and we have made it copy-and-paste simple. Any one of these is fine — pick whichever fits where you are using the data.
Plain text
Measured by batchwatch — https://batchwatch.devHTML
Measured by <a href="https://batchwatch.dev">batchwatch</a>Markdown
Measured by [batchwatch](https://batchwatch.dev)Citation line
batchwatch — measured LLM batch-API queue times. https://batchwatch.dev (CC BY 4.0).The required attribution is a credit to batchwatch and a link back to https://batchwatch.dev. If you have changed or reprocessed the numbers, say so, so a reader does not attribute your derivation to us.
The badge is attribution by construction
The easiest way to comply is the live status badge — it carries a link back to batchwatch by construction, so embedding it is attribution. If you want to show a current number and credit us in one move, use the badge and there is nothing else to do.
It is served in-band, too
Every open data endpoint returns the licence with the data, so a programmatic reuser gets the attribution requirement without reading this page. /v1/coverage and the other open endpoints carry a license block:
"license": {
"id": "CC-BY-4.0",
"url": "https://creativecommons.org/licenses/by/4.0/",
"attribution": "Measured by batchwatch - https://batchwatch.dev",
"attribution_html": "Measured by <a href=\"https://batchwatch.dev\">batchwatch</a>",
"source_url": "https://batchwatch.dev/license"
}The attribution string is ready to paste. Most people who omit a credit do so because working out the right form is friction, not out of malice — so we hand you the string and the link comes for free.
On enforcement — plainly
A licence on public data is a norm and a request, not a fence, and we will not pretend otherwise. We do not run detection and we do not send takedowns. If you use the numbers without crediting us, we treat that as a sign the data was useful, not as an infringement to chase. Crediting us is easy, it is the right thing, and it helps the next person find the source — that is the whole case for it.
Questions about reuse: hello@batchwatch.dev.