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.
Exhibit · side by side
| PARALLEL | GUMSHOE | |
|---|---|---|
| METHOD | Search + crawl of open web | Search + real browser navigation |
| JS WALLS / LOGINS / FILTERS | Out of reach | Driven like a human researcher |
| PROOF | Citations + reasoning per field | Citations + timestamped screenshots |
| LIST BUILDING | Not the core motion | Sweep: discover + dedupe + qualify |
| BEST FIT | Open-web deep research | Company research behind interaction |
| PRICING | Task 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 caseOther case files