The Stack Overflow API here is a thin, paid front door to the official Stack Exchange API — you send a keyword and/or a tag filter and get back one typed row per question: title, score, answer count, view count, the answered and accepted-answer flags, tags, the asker with their reputation, the asked and last-activity timestamps and the permalink.
Stack Exchange already publishes this data without a key, and for a few hundred calls you should use it directly. What you are paying us for is the network underneath: the public API throttles per IP and starts returning a backoff the moment you pull at dataset scale, and our rotating residential exits spread the load so a 100-question sweep does not stall on a single flagged address. You send your QuantumProxies key; there is no Stack Exchange registration to manage.
$0.0003 per delivered question, up to 100 questions per run. Nothing delivered means nothing charged.
Each run is one search against a Stack Exchange site (default stackoverflow, but any sister site works). You can pass a query, a semicolon-separated tags filter, or both, and choose a sort — votes, activity, creation or relevance.
answer_count and has_accepted_answer are what ranking and triage work actually need.Everything is optional except that you need at least a query or a tags filter to have something to search. site switches to any Stack Exchange property (serverfault, superuser, math…), sort orders the result set, and max_results caps delivery at 100 — you pay only for the questions that come back.
One flat row per question. score, answer_count and view_count are the popularity signals; is_answered and has_accepted_answer separate solved from open. tags is an array, author and author_reputation identify the asker, and link is the canonical permalink so you can join back to the live thread.
$0.0003 per delivered question ($0.30 per 1,000). Nothing delivered means nothing charged, and the $2 monthly free credit covers roughly 6,666 questions here. Volume tiers take up to 30% off.
$0.0003 per delivered question — $0.30 per 1,000. There is no subscription, and a query that matches nothing simply costs nothing. The free monthly allowance is $2 of usage with no card, which at this price is roughly 6,600 questions a month.
Because you are billed per delivered row rather than per request, a broad tag sweep that returns fewer questions than your max_results cap costs less than the cap, not more.
The Stack Exchange API is free, well documented and does not require a key for read access — so be honest with yourself about scale. For interactive or low-volume use, call it directly; you do not need us.
The collector earns its price in two situations. The first is volume: the public API meters aggressively per IP and per key, and a large historical pull will spend more time backing off than fetching unless the requests come from many addresses. The second is uniformity — this returns the same flat, typed schema as every other QuantumProxies collector, so Stack Overflow questions land in the same pipeline as your npm and CVE rows without a bespoke client per source.
Wrapping the Stack Exchange endpoint is an afternoon of work; keeping it fed at scale is the part that bites. You end up building an IP rotation layer to survive the throttle, a backoff handler, and a normaliser to turn the nested JSON and HTML entities into clean rows — which is exactly the collector.
If you already have that plumbing and only want cleaner exits under it, run your own client over our residential proxies instead.
Pull the top questions for a tag set sorted by activity to see what a technology's users are struggling with this month — view_count and answer_count tell you demand versus the supply of answers.
Feed a query and rank by score to find the canonical questions worth writing tutorials or docs against, then use link to cite them.
Watch a product's tag for new, still-unanswered questions (is_answered false) so your team can step in before a thread goes cold.
One run delivers up to 100 questions from the public search. Bigger corpora are more runs, which parallelise up to your plan's rate limit (60 requests/minute on pay-as-you-go, up to 1,200 on the top tier). Because we honour the source's own throttle, an extremely aggressive single job can still see individual calls deferred rather than dropped.
Stack Exchange content is licensed CC BY-SA and the API has its own terms; attribution and reuse rules apply to what you publish downstream. This is not legal advice — check the current terms for your use case.
The underlying Stack Exchange API is free and keyless for reads, so for small jobs use it directly. This collector adds a paid infrastructure layer — rotating exits and one uniform schema — and carries a $2/month free allowance of its own, about 6,600 questions, with no card.
No. You authenticate with your QuantumProxies key only. There is no Stack Exchange app registration, client id or access token to obtain or rotate.
Yes. Set the site input to any Stack Exchange property — serverfault, superuser, askubuntu, math and so on — and the search runs there instead. The default is stackoverflow.
No. A row covers the question and its metrics — score, answer count, accepted-answer flag, tags and author. Full answer bodies are a separate concern; the permalink in link lets you fetch a thread when you need the prose.