Gumshoe.Open a case

Case file · parallel · web research API

The Parallel alternative for research that needs to walk the real web.

Parallel is the best at reading and citing the open web. Gumshoe is built for what search can't reach: JS walls, logins, portals, filters: the sites you have to drive.

Start freeSee what a case returns

Exhibit · side by side

PARALLELGUMSHOE
METHODSearch + crawl of open webSearch + real browser navigation
JS WALLS / LOGINS / FILTERSOut of reachDriven like a human researcher
PROOFCitations + reasoning per fieldCitations + timestamped screenshots
LIST BUILDINGNot the core motionSweep: discover + dedupe + qualify
BEST FITOpen-web deep researchCompany research behind interaction
PRICINGTask API $5 to $2,400 per 1k requests by depth. Search $1 to $5/1k. No real browser navigation.$0.037 to $0.05 per company, browser fieldwork opt-in at 25 credits when a site needs driving.

THE WEDGE

Parallel is excellent at synthesizing citable answers from searchable content. But a lot of the truth about a company lives behind interaction: filtered directories, gated pricing, JS-heavy careers pages, login walls. That's where a search-and-crawl API stops and a real browser agent keeps going.

WHEN PARALLEL IS THE RIGHT CALL

If your research lives on the open, searchable web and you want best-in-class citations and calibrated confidence, Parallel is a superb choice, with real enterprise credibility.

Parallel: Web infrastructure for AI agents with a strong deep-research Task API and the Basis framework: citations, reasoning, and calibrated confidence per field. Serious, benchmark-leading.

Put a detective on it

100 free cases a month, no card. Send a question, get a closed case back: verdict, confidence, citations, screenshots.

Open your first case

Other case files