Documentation
The baba News API, in 13 pages
17 endpoints over the Israeli press: clustered stories with the newsrooms that filed and the ones that did not, running threads, coverage gaps, Hebrew reporting translated to English, newsroom profiles, search, the daily brief and named people.
Pages
MCP for AI agentsConnect your agent to 16 News API tools in Hebrew and English.QuickstartA key, one header, and the first request.AuthenticationThe bearer header, the key format, rotation, and what a rejection looks like.Scopes and plansWhich endpoints a key can reach, and what each plan carries.Rate limits and quotaThe per-minute limit, the monthly quota, and the two different 429s.ErrorsOne envelope, nine codes, and which of them are worth retrying.Paging and filtersKeyset cursors, the shared filters, conditional reads, and the story object.Endpoint referenceEvery path, parameter and response, generated from the OpenAPI document.WebhooksPush delivery of breaking and blindspot events, with signing and retries.Bulk exportHistorical articles and clusters as JSONL or CSV, up to 366 days a range.OpenAPI and SDKsThe machine-readable document, code generation, and how versions move.Attribution and supportWhat you must preserve when you display coverage, what the API is not, and how to reach us.Agent skillThe whole reference as a downloadable skill for a coding agent.
Every endpoint
The full parameter list, response shape and an example for each of these is on the endpoint reference.
endpoints.md
| Endpoint | Scope | What it returns |
| --- | --- | --- |
| `GET /api/v1/stories` | stories | List stories |
| `GET /api/v1/stories/{ckey}` | stories | Story detail: members, timeline, framing |
| `GET /api/v1/stories/{ckey}/members` | stories | Every article in a cluster, paged |
| `GET /api/v1/threads` | stories | List running stories |
| `GET /api/v1/threads/{id}` | stories | One running story and its whole arc |
| `GET /api/v1/trending` | stories | Newsroom momentum ranked by an explicit computed metric (newsroom-momentum-v1) |
| `GET /api/v1/blindspots` | stories | Coverage gaps, by camp |
| `GET /api/v1/sources` | sources | List newsrooms |
| `GET /api/v1/sources/{key}` | sources | One newsroom, with profile when computed |
| `GET /api/v1/sources/{key}/activity` | sources | Measured newsroom behaviour |
| `GET /api/v1/topics` | sources | The desk register with live volume |
| `GET /api/v1/changes` | stories | Incremental change feed |
| `GET /api/v1/search` | search | Search stories |
| `GET /api/v1/people` | people | People in the last 24 hours |
| `GET /api/v1/people/{id}` | people | One person, with every spelling folded into them |
| `GET /api/v1/brief` | brief | Today in Brief |
| `GET /api/v1/export` | export | Enterprise historical article and cluster export |
| `GET /api/v1/openapi.json` | — | This OpenAPI document |Markdown, everywhere
Every page above is served as raw Markdown at /api/developers/docs/<page>.md, and each page carries a button that copies its own source. The whole set, plus the OpenAPI document, downloads as an agent skill a coding agent can read before it writes any integration code.