How we score, in full

1. Data sourcing

Every factual field — pricing, EHR integrations, compliance, features — is taken from official vendor pages or reputable secondary coverage. Each tool lists its sources, a last-verified date and a data-confidence rating. Where a vendor does not publicly disclose something, we record it as “not disclosed” and never imply a “no”. Third-party price estimates are never presented as official pricing.

2. Hands-on testing vs. sourced facts

The two dimensions clinicians care about most — clinical precision and note quality — cannot be read off a vendor website. We score them by hand, after testing. The remaining three dimensions are derived from sourced public facts (“not disclosed” scores as neutral half-credit, never a penalised “no”). The weighting:

The data-derived dimensions are computed in lib/data/scoring.tsand reproducible from the public facts. For a tool we've tested, the overall score is the reviewers' considered judgement after hands-on use, led by clinical precision and note quality. A tool we haven't tested yet shows a clearly-labelled provisionalscore from public data only. That score is deliberately mapped into a band below the weakest tested tool — strong paperwork alone can't lift an unproven tool into the tested tier, so a provisional score and a tested score are never read as equivalent.

3. Editor's Pick (hands-on)

The computed score can only see what vendors publish. It can't measure the thing clinicians care about most: how good the note actually is, and how much manual editing it needs. Our Editor's Pick is a separate, clearly-labelled editorial designation from hands-on testing. It is pinned to the top of the ranking; every other tool below it stays strictly in computed-score order. The pick is opinion, the score is a formula, and we never present one as the other.

4. Verified reviews

Clinician reviews are separate from both the score and the Editor's Pick. Every review requires email verification before it's published, is rate-limited and de-duplicated per email, and can be removed if it breaches our guidelines. Aggregate star ratings are always shown distinctly so the three are never conflated.

5. Independence & disclosure

CompareScribes currently has no affiliate, commission or partnership deals with any vendor. No commercial deal influences anything on this site. The data-derived sub-scores are a deterministic function of public facts; the Editor's Pick is the single, disclosed editorial call. If we ever introduce a commercial relationship, it will be disclosed here first — and even then it would never affect scores, the Editor's Pick or ordering.

Editorial-team disclosure. Members of the editorial team may have prior or current professional relationships with tools in the catalog; rankings are score-driven and no commercial deals influence placement.

6. Corrections

Found something out of date or wrong? Tell us and we'll re-verify and update the source and date. Accuracy beats being first.