Try any of 3868 endpoints — live.
Pick an endpoint, load a working example, tweak the params, and send — no signup to try. Results render the way the data deserves; raw JSON, headers & code are one tab away.
Search live BUYER project requests on the Kwork exchange. Returns the full description, the buyer's budget cap in roubles, the deadline, the category, how many offers the request has already received, and the buyer (id, username, profile URL, avatar, projects posted, hire rate, purchase-level badges). No second call is needed: Kwork publishes no project detail page and every field is already here. Cost is bounded by `max_pages`, and the response states how many pages were read.
Free-text search, applied by Kwork itself across BOTH the title and the body of the request (measured: for 'telegram', 7 of the first 12 matches were in the title and 5 matched only in the body). The exchange is Russian-language, so Russian terms match far more than English ones.
Restrict to one Kwork category, by slug ('script-programming') or numeric id ('41'). Validated against the live category tree: Kwork itself does NOT reject an unknown category, it silently returns the whole exchange, so we reject it here instead of answering a narrow question with everything.
Keep only projects whose buyer budget is at least this many roubles. Kwork does not offer this over HTTP, so it is applied on our side to the pages covered by `max_pages` — see `budget_exhausted` in the response. (0–)
Keep only projects whose buyer budget is at most this many roubles. Applied on our side, like `budget_min`. (0–)
Keep only projects that have received at most this many offers, for finding requests that are still winnable. Of 657 open projects measured, 102 had no offers at all. Applied on our side. (0–1000)
How many projects to return (1–120). The page budget still applies, so a heavily filtered request can return fewer. (1–120)
Skip this many projects before returning `limit` (0–660; the exchange is about 55 pages of 12). (0–660)
Hard cap on how many exchange pages this one call may read (1–10, default 3). This is the cost control: the call reads at most this many pages whatever the filters, and the response reports `pages_read` and `budget_exhausted` so you can tell a truncated result from an exhausted one. (1–10)
curl -X POST https://api.reefapi.com/kwork/v1/search \
-H "x-api-key: $REEF_KEY" \
-H "content-type: application/json" \
-d '{"limit":"10"}'Press Send request to see the live result here.
Not sure what to send? Ask Reef AIWhat you can run here, without an account
ReefAPI is one HTTP gateway to 633 live web-data APIs covering 3868 endpoints across 13 categories. Every endpoint answers with the same JSON envelope, so one integration reaches all of them. Each response reports what it cost and when the source was read.
No signup, no API key and no card are required to try an endpoint on this page. Anonymous calls run on a shared demo key, capped at 15 calls per hour per visitor. Pick an API, fill the parameters and the real response comes back from the live source, not from a canned sample. A free account lifts the cap and gives you 1,000 credits and your own key.
Browse every API in the catalog, read an endpoint's fields and example response in the docs, or see per-call credit costs on the pricing page.
Coverage by category
| ecommerce | 350 |
| classifieds | 117 |
| real-estate | 29 |
| media | 25 |
| utilities | 18 |
| jobs | 18 |
| developer-tools | 18 |
| reviews | 15 |
| social | 15 |
| search-seo | 11 |
| travel | 8 |
| other | 5 |
| finance | 4 |