How it works, what it costs, and where the boundaries are

You describe the data you need and your must-have conditions. Our search agent runs a real search across open data platforms, academic and industry sources and institutional sites, judges each candidate against your conditions, and delivers an availability assessment report. Unlike searching yourself, it covers many sources at once, judges item by item, cites a traceable source for every result — and states gaps honestly instead of forcing a match.

Sign-up includes 3 free credits, good for 3 standard searches. After that each standard search costs 1 credit (1 credit = 10 CNY). Credits never expire. Viewing the report — dataset names, source links and the item-by-item judgement — requires a signed-in account.

A deep search costs 10 credits. On top of the standard search it expands more proper names, abbreviations and bilingual aliases, runs more search paths and sources, and verifies candidates more strictly, producing a more professional and comprehensive report. It suits cases where the data should exist but a standard search did not surface it. Deep search is billed per run and does not guarantee a match.

Searches are billed per run, not per hit. Once the report is generated and viewable, the search is complete; a no-match is not refunded — the real search process and candidate list are the deliverable. If a system failure prevents the report from being generated or viewed, or the report clearly falls short of the service description, we refund or compensate the credits.

Typically candidate sources, an item-by-item match judgement against your conditions, a gap analysis and next-step suggestions. The search process is transparent: which queries were used, what each source returned, and which candidates were excluded and why.

No. The assessment delivers source links, the search process and the judgements — we do not download third-party raw data on your behalf. Data is obtained under each source's own licence and application path, which the report spells out.

We assess availability first, verifying sources, licences and access paths one by one. If the data truly cannot be obtained, we say so plainly and propose feasible alternatives rather than over-promising.

Real collection comes first. If certain indicators cannot be obtained directly and you accept model-based filling, we clearly separate collected from estimated parts and note the estimation method and share — estimates are never passed off as real data.

Client identity and commissioned work are confidential by default. Any samples shown publicly are anonymised with the client removed.

No. Credits have no expiry date.