OpenSERP Cloud is the hosted version of OpenSERP. You get the same multi-engine SERP API as the open-source server - Google, Bing, Yandex, Baidu, DuckDuckGo, Ecosia, plus megasearch - without running browsers, proxies, or queues yourself.
Already running OSS? Cloud uses the same response schema. Switching is changing the base URL and adding an
Authorizationheader.
Who Cloud is for
| You’re doing… | Cloud gives you… |
|---|---|
| Building a product or AI agent that needs live SERP data | A single REST endpoint, 99.9% target uptime, no scraper to maintain |
| Tracking rankings or competitor listings on a schedule | Predictable per-call billing, no infra spend, no captcha babysitting |
| Grounding LLM responses on fresh search results | Multi-engine search, megasearch, and fallback routing in one API |
| Prototyping or doing one-off research | Pay-as-you-go credits, no monthly minimum |
If you’d rather host it yourself - for compliance, custom proxy setups, or zero per-call cost - the open-source server is production-ready and free under MIT.
How Cloud differs from OSS
| OpenSERP OSS | OpenSERP Cloud | |
|---|---|---|
| Hosting | You run it (Docker, Go binary) | We run it |
| Proxies | You bring your own | Managed pool |
| CAPTCHA / blocking | You handle retries | Managed |
| Scaling | You provision | Elastic |
| Auth | None / your own | API key, per-account |
| Pricing | Free (self-hosted infra cost) | Pay-as-you-go credits |
| Schema | Same JSON envelope | Same JSON envelope |
The endpoint paths and response shapes are intentionally close to the OSS server. Code that works against an OSS server usually needs only the Cloud base URL and an Authorization header.
What you can do today
- Search any single engine -
/v1/google/search,/v1/bing/search, etc. See Endpoints. - Megasearch across engines -
/v1/mega/searchfor merged, deduplicated coverage. - Any/Fast routing -
/v1/any/searchand/v1/fast/searchfor 1-credit fallback search. - Image search - single-engine, aggregated, or Any/Fast image fallback.
- Filters - language, country, date range, site, file type, pagination.
- Track usage -
X-Credits-UsedandX-Credits-Remainingon every billable response.
On the roadmap
These are landing in the next few releases - your feedback shapes the order:
- Scheduled / recurring searches with delivery to webhook or storage
- AI extraction (entities, summaries, structured fields) layered on results
- Web archive snapshots
- Per-team API keys and usage breakdowns
- SDKs for Node, Python, and Go
If you need one of these now, tell us at [email protected] - we prioritize what early users actually use.
Start here
- Quickstart - first request in five minutes.
- Authentication - keys, rotation, header format.
- Endpoints reference - every endpoint and parameter.
- Pricing & credits - what calls cost and how billing works.
- Errors & rate limits - status codes and retry guidance.
Or skip the docs and try the Search Playground - paste a query, run it against any engine, and copy a working curl from the live request preview.