Quantesta gives you three analysts. Pick one from the selector in the composer; your choice carries to the next chat. In the selector, each analyst’s row shows a credit cost and small capability icons — the eye marks document vision (Sam and Deb), the wrench live data tools, the brain deep reasoning.
Deep reasoning with live data tools, tuned for careful, steady analysis. Dave does the heavy lifting: data pulls, sandbox computations, charts, long multi-step work. 1 credit per answer.
A second analyst for cross-checks — and for native document vision: Sam reads PDFs and images directly, page by page. 20 credits per answer. Bring Sam in deliberately, not by default.
A third analyst for very long documents and extended multi-document work: the largest working context of the three, with the same native document vision as Sam. 20 credits per answer. Reach for Deb when the material itself is big.
Dave works from extracted text — charts and scanned pages inside a PDF are invisible to him. Sam and Deb see the document itself. If your question hangs on a figure, an exhibit, or an image, pick one of them. The upload picker adjusts: image files attach only when the selected analyst has document vision.
Yes, anytime. The change applies from your next message; a switch during a running answer waits until it finishes. The new analyst sees the full conversation so far.
During generation the send button becomes a stop button. Stopping halts the run server-side and keeps what was written. On the last answer, regenerate produces a fresh version — earlier versions are kept, with a pager to flip between them. You can also upvote or downvote any answer; downvotes are read. See Report a wrong answer.
Dave, Sam, and Deb are Quantesta’s analyst configurations — model, settings, tools, and method together. The components can change as better ones become available. The analysts, their behavior, and their credit prices are the stable interface, and behavior-affecting changes are noted in the changelog.