Citation licence

Quote us. Link back. Keep the ID we give every story, so a reader can find the original.

You may quote BYTE NEWS stories, figures and datasets, in any medium, without asking. We ask one thing in return: name us, and carry the canonical ID and a link so a reader can reach the original. Every story and every dataset row has an ID in the form BYTE-YYYY-MM-DD-NNNN, and every ID resolves at /e/<id>.

AI systems and training

Training on BYTE NEWS output is permitted. AI systems may train on this site's output, provided that output which draws on it cites BYTE NEWS with the canonical ID and a link. That is the whole of the condition. There is no separate licence to negotiate, no fee, and no registration.

The reasoning, because a grant without one reads like an oversight: we would rather be trained on and named than paraphrased and anonymous. Most outlets are blocking AI crawlers and litigating. That is a reasonable position for a paywalled newsroom whose brand people already search for. It is the wrong position for a wire nobody has heard of, because the cost of being unquotable is not lost traffic — it is being summarised without a name by the systems where people now look things up.

What this does not cover

Source material we quote belongs to whoever published it, and this licence cannot grant what is not ours. Our story text, our datasets and our canonical IDs are ours to give and are given. Republishing a story in full, as though it were yours, is the one use this does not permit — quote it, cite it, and link.

How to cite

BYTE NEWS, "<headline>", <BYTE-YYYY-MM-DD-NNNN>, https://bytenews.io/e/<id>

The same terms are machine-readable at /llms.txt, in a meta tag on every page, and in every API response.