Hebrew news API
Read the Hebrew press in English, over HTTPS.
39 Hebrew-language Israeli newsrooms, scraped continuously, translated into English, and clustered by event so you can see which part of the press filed and which part stayed quiet.
What it returns
A request to /api/v1/stories with lang=en returns clustered events. Each cluster names the newsrooms that covered it, the side of the press each one sits on, and the state of the story: breaking, developing, or not-everywhere when a side of the press has not filed. Headlines, summaries and bodies arrive in English whether the newsroom published in Hebrew, Arabic, Russian or French. Pass lang=he and the same objects come back in Hebrew.
curl --get 'https://news.itsbaba.com/api/v1/stories' \ --header "Authorization: Bearer $BABA_NEWS_API_KEY" \ --data-urlencode 'lang=en' \ --data-urlencode 'state=blindspot' \ --data-urlencode 'limit=20'
The response carries data, nextCursor, lang and generatedAt. Story IDs identify clusters, article IDs identify individual reports, and both are stable. The endpoint reference lists every filter and field; the guides walk the first request, pagination, webhooks and export.
The newsrooms behind it
The register is fixed and published, which is what makes an absence measurable. All 59 sources are listed at /sources with owner, funding, place and founding year, and the same records are readable through /api/v1/sources.
| Publishing language | Newsrooms |
|---|---|
| Hebrew | 39 |
| English | 7 |
| Arabic | 6 |
| Russian | 5 |
| French | 1 |
| de | 1 |
By side of the press the register holds 17 newsrooms rated center, 14 right and 4 left, with 24 unrated because a business paper, a public broadcaster or a government wire does not sit on that axis and pretending it does would make the split worse, not better.
How it differs from a country filter
Every general news API sells Israel as a value for a country parameter, and returns the Hebrew as Hebrew. Two things follow from that, and both are the reason this endpoint exists.
| Question | A general news API | This API |
|---|---|---|
| What language do I get? | The publishing language. Hebrew stays Hebrew. | English or Hebrew, from the same records. The translation is the product. |
| What is the unit? | An article. | An event, with every article that covered it attached. |
| Who covered it? | Whoever the crawl happened to reach. | A fixed register of 59 sources, so the list is comparable across days. |
| Who did not cover it? | Not answerable. There is no register to measure against. | Returned per story, as the sides missing from the cluster. |
| Where does the outlet sit? | Not modeled. | Lean, ownership, funding and median filing lag per newsroom. |
What it costs
Starter is $99 a month for 10,000 requests and 60 a minute, covering stories, the newsroom register and search. Team is $299 a month for 100,000 requests and 300 a minute, adding the daily brief and people. Enterprise adds signed webhooks and bulk export under a written agreement and is priced on request. Every plan opens with a 14-day trial: one key, 10,000 requests, Team read scopes, no card. Every limit, side by side.
Questions
Does the Hebrew news API return Hebrew or English?
Both. Pass lang=en and every headline, summary and body arrives translated into English. Pass lang=he and you get the Hebrew. The English is a translation the service produces, not an English edition an outlet happened to publish, so Hebrew-only newsrooms are readable either way. Where a translation is not yet available the Hebrew response falls back to English.
How many Hebrew newsrooms does it cover?
39 of the 59 monitored sources publish in Hebrew. The rest publish in English (7), Arabic (6), Russian (5) and French (1). Every one of them is listed at /sources with its owner, funding and place.
How is this different from a general news API with a country filter?
A country filter returns articles. This returns events. Articles covering one event are clustered, and the cluster carries the left, center and right split of the newsrooms in it plus the sides that did not file at all. That last part is the one a country filter cannot give you: it needs a fixed register of outlets to notice an absence, and a general API does not have one for Israel.
Can I get historical Hebrew coverage in bulk?
Yes, on the Enterprise plan, through /api/v1/export as CSV or JSONL over a date range of up to 366 days. Exports carry IDs, source, language, headline, summary, topic, timestamps, cluster state and the source-linked URL. They do not carry the publishers’ full article text.
What does the Hebrew news API cost?
Starter is $99 a month for 10,000 requests and covers stories, sources and search. Team is $299 a month for 100,000 requests and adds the daily brief and people. Enterprise adds webhooks and bulk export under a written agreement and is priced on request. Every plan starts with a 14-day trial that needs no card.
Do I have to credit the original newsroom?
Yes. Every record carries the originating source and a link back to it, and the terms require you to preserve both when you display coverage. The press-side labels and machine-written summaries are editorial signals, not factuality scores, and they are not a license to republish a publisher’s article.