Search Guru.com freelance jobs with budgets and quote counts

No login is involved. The engine's meta reports pii_redacted true and client_country came back null on the rows we read — you get the job and its terms, not the buyer.

Against us. total_count was null on every call we made — keyword, category, deep offset, all of them. There is no result count anywhere in the response, so you cannot show a user how many jobs exist and you cannot compute progress through a crawl. The workable substitute is asking for more than you expect: one skill returned 15 rows against a limit of 80, and one category returned 61 against the same limit, which is a reliable end signal.

The meta returns search_mode, resolved_skill and server_side. A keyword that maps onto Guru's own taxonomy comes back as canonical_skill with the resolved slug and server_side true — meaning the source filtered, not us. A category browse comes back as browse with server_side false. So you always know whether you are reading a real server-side filter or a local one, which is exactly the distinction that silently ruins keyword coverage elsewhere.

offset 500 on a skill that only has fifteen live jobs returned zero rows with stop_reason complete. There is no artificial cap being hit here — the board scope is simply small, and the engine says complete rather than pretending there is more.

budget comes back as an object with type, min, max and currency, and for hourly jobs a label written the way Guru writes it — hourly, a rate band, the weekly hours and the expected engagement length in one string. Fixed jobs carry the same object with a fixed type. The rate and the commitment arrive together.

Against us, and counter-intuitive. On the job we opened, search reported 49 quotes received, a category and a status; the detail for the same id returned bids_count, category and status all null. What detail does add is valid_through and employment_type. Read the grid for competition, the detail for the deadline — not the other way round.

posted_at is a Unix timestamp and posted_relative is the board's own line — "Posted on Aug 23, 2026 · 49 Quotes Received" — which carries the date and the competition in one string. There is no freshness filter parameter, so recency is something you sort on rather than ask for.

project takes the full job URL or the slug/id path that search returns on every row. A bare numeric id is not enough on its own. Since job ids on a bid board expire, resolve from a fresh search row rather than from a stored id.

The job and its terms: title, description, budget, skills, category, quote count, urgency and NDA flags, and the board link — with the client's identity stripped by the engine before you see it. A public provider-profile action also exists, returning what Guru itself shows on a provider page; we do not build this page around it and we do not sell profiles as a product. Nothing here carries a private contact channel.