Deep research you can check, line by line.
Every sentence is tied to a primary source with a provenance token and checked by an independent two-family evidence pipeline. Ask a question — get a brief where you can click any claim and read the exact passage behind it.
The estimated treatment differences were −0.15 percentage points (95% CI −0.28 to −0.03), −0.39 percentage points (95% CI −0.51 to −0.26), and −0.45 percentage points (95% CI −0.57 to −0.32), all favoring tirzepatide .
From a POLARIS brief answering: What is the efficacy and safety of tirzepatide for glycemic control and weight loss in adults with type 2 diabetes?
… at a dose of 5 mg, 10 mg, and 15 mg were −2.01 percentage points, −2.24 percentage points, and −2.30 percentage points, respectively, as compared with −1.86 percentage points with semaglutide (Figure 1A). All tirzepatide doses were found to be superior to semaglutide, with estimated treatment differences of −0.15 percentage points (95% confidence interval [CI], −0.28 to −0.03; P=0.02) with tirzepatide 5 mg, −0.39 percentage points (95% CI, −0.51 to −0.26; P<0.001) with tirzepatide 10 mg, and −0.45 percentage points (95% CI, −0.57 to −0.32; P<0.001) with tirzepatide 15 mg. The glycated hemoglobin level over time is shown in (Figure 1B). A total of 82 to 86% of the …
Highlighted text is the exact span cited by the claim.
Tirzepatide versus Semaglutide Once Weekly in Patients with Type ...
Provable
Click any claim and see the exact source passage it came from. Frontier tools hallucinate 3–13% of their citations — every POLARIS claim is span-anchored to a primary source.
Canadian-hosted
Built for Canadian-hosted deployment. Public sources are fetched via logged Canadian egress, and every brief is integrity-hashed and auditable.
Snowball
Each run grows a connected knowledge graph of claims and sources you can explore, follow up on, and build the next question from.