New Zealand · Pre-release
Who reported the announcement.
BeeLine indexes government press releases from beehive.govt.nz and shows which news outlets covered each one. Its subject is not the news but the sourcing of the news: what was announced, who carried it, and how closely their account tracks the release.
Every match is a headline and a link that sends the reader to the publisher.
Reads the Beehive
Every release published to beehive.govt.nz, taken from the Beehive's own RSS feed. Permission to use that feed for this project was requested and confirmed by email before anything was built on it.
Summarises, and grades its own summary
Each release gets a short machine-written summary, and each claim in that summary is checked back against the release text.
Finds the coverage
Where a publisher's own feed carries a story about the same announcement, BeeLine shows it as a headline and a link out. Often there is none, and the release says so plainly rather than filling the space.
Transport
Government confirms funding for regional rail upgrades
Three of four claims appear word-for-word in the release; one is paraphrased.
2 outlets covered this
The Spinoff
What the rail announcement actually funds, and what it doesn't
The Conversation NZ
Regional rail has been promised before — the record is mixed
What BeeLine asks for, and what it does with it
If you publish New Zealand news, this is the whole of the arrangement — stated here so you can check it against what BeeLine actually does rather than take it on trust.
- Article text, where it is licensed, is used to compute a similarity match between a government release and your coverage of it. That is its entire purpose.
- Readers are never shown your article text. A match displays as outlet, headline, time and a link that sends the reader to your site.
- Nothing is used to train or fine-tune a model. Matching uses text embeddings for retrieval, and no article text is retained in any model.
- Where full text is not licensed, BeeLine works from the headline, link, timestamp and short description a feed already publishes.
BeeLine is pre-release: no users and no revenue. A reader subscription is possible eventually.
BeeLineCoverageBot
If you have seen this user agent in your logs, it is BeeLine. It reads publisher RSS feeds to match press coverage to government releases.
- User agent
BeeLineCoverageBot/1.0- What it reads
- Publisher-operated RSS and Atom feeds only. Fetching full article pages is disabled in the code, behind a setting that exists to be switched on for publishers who have licensed it.
- robots.txt
- Checked before every feed fetch, and honoured, including crawl-delay. A feed that robots.txt disallows is skipped and logged.
- Rate
- A feed poll on a fixed cooldown. BeeLine does not crawl your site, follow links within it, or request article pages.
- To block it
- Add the rule below, or email the address at the foot of this page and it will be removed from the source list directly.
User-agent: BeeLineCoverageBot
Disallow: /There is a person at this address
Matthew Lee. Licensing enquiries, corrections, takedown requests and crawler complaints all reach the same inbox, and are answered by the person who wrote the code.