# Search everywhere (/docs/engines/search-everywhere) Send one query to many platforms at once and get one ranked, de-duplicated list, with the price shown free first Search everywhere sends one query to 18 platforms at once. It returns one list of posts, ranked and with near-duplicates collapsed. It is the quickest way to see what people say about a name, a brand or a topic. A dry run shows the plan and the price for 0 credits. ## Price it for free Add `dry_run=1` to a search. It runs nothing. ```bash title="cURL" curl "https://www.socialcrawl.dev/v1/search/everywhere?query=Liquid%20Death&period=7d&format=compact&dry_run=1" \ -H "x-api-key: YOUR_API_KEY" ``` ```json title="Response (trimmed)" { "success": true, "data": { "valid": true, "estimate": { "pricing": "flat", "hold": 20, "credits": { "min": 0, "expected": 20, "max": 20 }, "seconds": { "expected": 13, "max": 16 }, "formula": "20 credits per search." }, "plan": { "searches": 1, "queries": ["Liquid Death"], "window": { "period": "7d", "since": "2026-11-12", "until": "2026-11-18" } } }, "credits_used": 0 } ``` `estimate.credits` is the price. `plan` lists the platforms the search will reach, with the date window each is sent. A search that finds no rows anywhere is refunded, so the charge can be lower than `expected`. ## Run it Send the same request without `dry_run`. `format=compact` returns about 20 rows in about 20 KB, which suits an agent. `deadline_s` is how long the call waits before it hands you a run to follow. ```bash title="cURL" curl "https://www.socialcrawl.dev/v1/search/everywhere?query=Liquid%20Death&period=7d&format=compact&limit=20&deadline_s=25" \ -H "x-api-key: YOUR_API_KEY" ``` ```json title="Response (trimmed)" { "success": true, "data": { "items": [ { "id": "tt_b9v0xwhb4yvf", "platform": "tiktok", "record_type": "video", "url": "https://www.tiktok.com/@.../video/...", "likes": 1223, "views": 9347, "comments": 10, "published_at": "2026-10-27", "why": { "found_by": ["tiktok"], "matched_keywords": ["liquid death"] } } ], "run": { "run_id": "run_...", "status": "completed", "credits": { "spent": 20 } }, "next": { "watermark": "..." } }, "pagination": { "next_cursor": "...", "has_more": true }, "credits_used": 20 } ``` ## What does it cost? A search costs 20 credits, however many platforms answer. Each item in `keywords` is its own search, so three keywords cost 60. Paging through the stored run, reading it again and an answer from the shared cache cost 0. A search that finds nothing is refunded, and one where most platforms failed is half refunded. ## How do you follow it over time? - **Page for free.** Every search is stored as a run. `GET /v1/search/everywhere/runs/{run_id}` returns the next page with `cursor`. - **Only what is new.** Send `since_watermark` with the `next.watermark` of your last answer and rows you already received are left out. - **A watch.** `POST /v1/search/everywhere` with a `schedule` repeats the search every `1h`, `6h`, `12h`, `1d`, `7d` or `30d` and delivers only rows it has not delivered. Send `dry_run` first. It shows `credits_per_30_days` for the watch. A watch must follow a topic, a brand, an organisation or a person in a public or professional role. It must not follow a private individual. ## Next steps - [Universal social search](/docs/search/everywhere.md): Every parameter, the stored runs, lists of searches, watches and streaming. - [Discover](/docs/engines/discover.md): When you need depth, a date window and a judgment on every row. - [Events feed](/docs/events.md): Read what a watch found, in order, from a cursor.